Hi, I have to questions around what that command is doing. 1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not. 2). When it does list repositories between the two messages, what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
regards, Steve
On Sun, 24 Nov 2024 at 22:21, Stephen Morris steve.morris.au@gmail.com wrote:
1). Why is it producing the messages "Updating and loadingrepositories:" and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
This is what I see on a clean f41 Vagrant box:
[root@localhost ~]# yum repolist repo id repo name
fedora Fedora 41 - x86_64 fedora-cisco-openh264 Fedora 41 openh264 (From Cisco) - x86_64 updates Fedora 41 - x86_64 - Updates
[root@localhost ~]# dnf check-upgrade Updating and loading repositories: Fedora 41 - x86_64 - Updates 100% | 6.9 MiB/s | 5.5 MiB | 00m01s Fedora 41 - x86_64 100% | 21.0 MiB/s | 35.4 MiB | 00m02s Fedora 41 openh264 (From Cisco) - x86_64 100% | 4.3 KiB/s | 6.0 KiB | 00m01s Repositories loaded. <snip bunch of packages that are liable for upgrade>
I would suspect that "Updating and loading repositories:" is checking the upstream repo metadata (if you've ever managed your own repos, this is the data created/updated by createrepo based on the RPM contents in the repo directory.)
This is the stuff you'll see like: https://fedora.mirrorservice.org/fedora/linux/development/41/Everything/x86_... those archives contain a lot of the pre-calculated RPM package dependencies (the requires/provides.)
Your local dnf will have its own cache of this metadata (unless this has changed significantly for dnf5) which is refreshed by default every 90 minutes. `dnf makecache` creates/updates this cache, `dnf clean` blats it causing subsequent runs to pull a new version of the cache.
On 25/11/24 09:58, Will McDonald wrote:
On Sun, 24 Nov 2024 at 22:21, Stephen Morris steve.morris.au@gmail.com wrote:
1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.This is what I see on a clean f41 Vagrant box:
[root@localhost ~]# yum repolist repo id repo name fedora Fedora 41 - x86_64 fedora-cisco-openh264 Fedora 41 openh264 (From Cisco) - x86_64 updates Fedora 41 - x86_64 - Updates
[root@localhost ~]# dnf check-upgrade Updating and loading repositories: Fedora 41 - x86_64 - Updates 100% | 6.9 MiB/s | 5.5 MiB | 00m01s Fedora 41 - x86_64 100% | 21.0 MiB/s | 35.4 MiB | 00m02s Fedora 41 openh264 (From Cisco) - x86_64 100% | 4.3 KiB/s | 6.0 KiB | 00m01s Repositories loaded.
<snip bunch of packages that are liable for upgrade>
I would suspect that "Updating and loading repositories:" is checking the upstream repo metadata (if you've ever managed your own repos, this is the data created/updated by createrepo based on the RPM contents in the repo directory.)
This is the stuff you'll see like: https://fedora.mirrorservice.org/fedora/linux/development/41/Everything/x86_... those archives contain a lot of the pre-calculated RPM package dependencies (the requires/provides.)
Your local dnf will have its own cache of this metadata (unless this has changed significantly for dnf5) which is refreshed by default every 90 minutes. `dnf makecache` creates/updates this cache, `dnf clean` blats it causing subsequent runs to pull a new version of the cache.
Yes, but I would expect those messages to only be produced in the situation you have shown, where it actually did find something to do. If there is nothing to do I would expect those two messages to not be displayed and a message to the effect that the local environment was up to date.
regards, Steve
On Mon, 25 Nov 2024 at 21:46, Stephen Morris steve.morris.au@gmail.com wrote:
On 25/11/24 09:58, Will McDonald wrote:
On Sun, 24 Nov 2024 at 22:21, Stephen Morris steve.morris.au@gmail.com wrote:
1). Why is it producing the messages "Updating and loadingrepositories:" and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
I would suspect that "Updating and loading repositories:" is checking the upstream repo metadata (if you've ever managed your own repos, this is the data created/updated by createrepo based on the RPM contents in the repo directory.)
Yes, but I would expect those messages to only be produced in the situation you have shown, where it actually did find something to do. If there is nothing to do I would expect those two messages to not be displayed and a message to the effect that the local environment was up to date.
So if I understand you then, your primary concern here is the semantic accuracy of the messages?
Updating and loading repositories Repositories loaded.
And you would rather something like:
Scenario: DNF Cache Outdated
Checking repositories cache status Cache out-of-date, updating from upstream repositories Repository cache updated
Scenario: DNF Cache Current
Checking repositories cache status Cache current, nothing to see here.
(Or whatever.)
You might want to take this to the upstream project, maybe submitting a bug report, documentation improvement, or PR if you feel particularly invested?
https://github.com/rpm-software-management/dnf5/discussions https://github.com/rpm-software-management/dnf5/issues
On 26/11/24 09:05, Will McDonald wrote:
On Mon, 25 Nov 2024 at 21:46, Stephen Morris steve.morris.au@gmail.com wrote:
On 25/11/24 09:58, Will McDonald wrote:On Sun, 24 Nov 2024 at 22:21, Stephen Morris <steve.morris.au@gmail.com> wrote: 1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not. I would suspect that "Updating and loading repositories:" is checking the upstream repo metadata (if you've ever managed your own repos, this is the data created/updated by createrepo based on the RPM contents in the repo directory.)Yes, but I would expect those messages to only be produced in the situation you have shown, where it actually did find something to do. If there is nothing to do I would expect those two messages to not be displayed and a message to the effect that the local environment was up to date.So if I understand you then, your primary concern here is the semantic accuracy of the messages?
Updating and loading repositories Repositories loaded.
And you would rather something like:
Scenario: DNF Cache Outdated
Checking repositories cache status Cache out-of-date, updating from upstream repositories Repository cache updated
Scenario: DNF Cache Current
Checking repositories cache status Cache current, nothing to see here.
(Or whatever.)
You might want to take this to the upstream project, maybe submitting a bug report, documentation improvement, or PR if you feel particularly invested?
https://github.com/rpm-software-management/dnf5/discussions https://github.com/rpm-software-management/dnf5/issues
Thanks Will, I will consider doing that. I was more looking for a message like "Repositories up to date, nothing to do" as other software I've used does. With the explanation of what dnf is doing, if an end used runs the dnf clean command with whatever obligatory sub-command they chose will their repositories lose all knowledge of what packages have been updated, or will dnf update/upgrade refresh that data irrespective of whether --refresh is specified or not?
regards, Steve
On Mon, 25 Nov 2024 at 22:23, Stephen Morris steve.morris.au@gmail.com wrote:
With the explanation of what dnf is doing, if an end used runs the dnf clean command with whatever obligatory sub-command they chose will their repositories lose all knowledge of what packages have been updated, or will dnf update/upgrade refresh that data irrespective of whether --refresh is specified or not?
I don't want to assume on your behalf. But you might be conflating/confusing the repository metadata with the DNF's local state and/or the RPM database history.
If you've only ever used DNF / Yum to manage boxes without needing to look at what RPM does I can see why you might worry. Nowadays it's probably very easy to manage systems with dnf/yum and not really have to dig into their interface into the underlying RPM stuff.
Repo metadata is this stuff: https://blog.packagecloud.io/yum-repository-internals/
It's hosted remotely with the repositories, mirrored locally for speed when cache expires and is pretty much entirely throwaway, because you can (normally) always get a refreshed copy from the remote repository.
"knowledge of what packages have been updated" is open to interpretation. What's been updated/available *in the upstream repositories* is entirely in that "remote" repository metadata.
What's been updated *on a local system,* is separate, distinct. I would guess that the source of truth here would be the rpmdb, maybe with a scattering of DNF data sources for speed, I've not looked into it in any detail.
The dnf5 man page states:
FILES Cache Files /var/cache/libdnf5/
Main Configuration /etc/dnf/dnf.conf
Repository Configuration /etc/yum.repos.d/
Repository Persistence /var/lib/dnf/
System State /usr/lib/sysimage/libdnf5/
I would hazard a guess that the DNF history is coming from System State?
[vagrant@localhost ~]$ tree /usr/lib/sysimage/ /usr/lib/sysimage/ ├── libdnf5 │ ├── comps_groups │ │ ├── cloud-bootloader-tools.xml │ │ ├── cloud-server-environment.xml │ │ ├── cloud-server.xml │ │ ├── container-management.xml │ │ └── core.xml │ ├── environments.toml │ ├── groups.toml │ ├── modules.toml │ ├── nevras.toml │ ├── packages.toml │ ├── system.toml │ ├── transaction_history.sqlite │ ├── transaction_history.sqlite-shm │ └── transaction_history.sqlite-wal └── rpm ├── rpmdb.sqlite ├── rpmdb.sqlite-shm └── rpmdb.sqlite-wal
4 directories, 17 files
I'd go poke around in these locations (in a test VM or container), read some man pages and maybe skim some of the project docs to understand more if it's something you feel you really need to know inside out? Possibly (and again, apologies if you already have, don't want to assume) but if you haven't done much directly with RPM, mess about with that to understand the separation between RPM and Yum/DNF?
On Mon, 2024-11-25 at 22:47 +0000, Will McDonald wrote:
What's been updated *on a local system,* is separate, distinct. I would guess that the source of truth here would be the rpmdb, maybe with a scattering of DNF data sources for speed, I've not looked into it in any detail.
I was under the impression that info about what's installed on your system is still RPM databasing, with DNF history essentially just being a log of DNF transactions.
On 11/25/24 5:02 PM, Tim via users wrote:
On Mon, 2024-11-25 at 22:47 +0000, Will McDonald wrote:
What's been updated *on a local system,* is separate, distinct. I would guess that the source of truth here would be the rpmdb, maybe with a scattering of DNF data sources for speed, I've not looked into it in any detail.
I was under the impression that info about what's installed on your system is still RPM databasing, with DNF history essentially just being a log of DNF transactions.
That is correct.
On 26/11/24 09:47, Will McDonald wrote:
Repo metadata is this stuff: https://blog.packagecloud.io/yum-repository-internals/
It's hosted remotely with the repositories, mirrored locally for speed when cache expires and is pretty much entirely throwaway, because you can (normally) always get a refreshed copy from the remote repository.
"knowledge of what packages have been updated" is open to interpretation. What's been updated/available *in the upstream repositories* is entirely in that "remote" repository metadata.
What's been updated *on a local system,* is separate, distinct. I would guess that the source of truth here would be the rpmdb, maybe with a scattering of DNF data sources for speed, I've not looked into it in any detail.
My main concern was, if I use dnf to install a local rpm and afterwards I issue sudo dnf clean all to reclaim space (not that at the moment I have a need to be concerned about disk space) does that remove knowledge of the rpm having been installed? My assumption is no, given that the remote repositories packages are rpms anyway, but I didn't know whether remote rpms and local rpms are handled differently because of going through a different internal code path and hence having different restriction/requirements. Also in the past I have people on this mailing list query why I was using RPM rather than, at the time, YUM, and that I should be doing the installs through YUM.
regards, Steve
On Nov 26, 2024, at 16:30, Stephen Morris steve.morris.au@gmail.com wrote:
My main concern was, if I use dnf to install a local rpm and afterwards I issue sudo dnf clean all to reclaim space (not that at the moment I have a need to be concerned about disk space) does that remove knowledge of the rpm having been installed? My assumption is no, given that the remote repositories packages are rpms anyway, but I didn't know whether remote rpms and local rpms are handled differently because of going through a different internal code path and hence having different restriction/requirements. Also in the past I have people on this mailing list query why I was using RPM rather than, at the time, YUM, and that I should be doing the installs through YUM.
The rpm database keeps track of what packages are installed, but it doesn’t know general repo information (such as what is not installed), that is part of the dnf (formerly yum) database. Dnf keeps track of where you installed a package from, and what happened in that transaction. People recommend using dnf instead of rpm because dnf does all the hard work of identifying all the dependencies and downloading anything else that’s needed, and importantly, validating the signatures and performing a test transaction before installing.
Neither of these databases are cleared when you run “dnf clean all”. That operation only removes all metadata and downloaded package files, things that can be re-downloaded on demand.
There are no “remote rpms” vs “local rpms”. You might say an RPM is installed, but the file used to install the package is not kept on the system, the files are extracted and placed on the filesystem, scripts are run, and the metadata is written to the database. After that, dnf deletes the downloaded package files from the cache, it is no longer needed.
On 27/11/24 08:50, Jonathan Billings wrote:
On Nov 26, 2024, at 16:30, Stephen Morrissteve.morris.au@gmail.com wrote:
My main concern was, if I use dnf to install a local rpm and afterwards I issue sudo dnf clean all to reclaim space (not that at the moment I have a need to be concerned about disk space) does that remove knowledge of the rpm having been installed? My assumption is no, given that the remote repositories packages are rpms anyway, but I didn't know whether remote rpms and local rpms are handled differently because of going through a different internal code path and hence having different restriction/requirements. Also in the past I have people on this mailing list query why I was using RPM rather than, at the time, YUM, and that I should be doing the installs through YUM.
The rpm database keeps track of what packages are installed, but it doesn’t know general repo information (such as what is not installed), that is part of the dnf (formerly yum) database. Dnf keeps track of where you installed a package from, and what happened in that transaction. People recommend using dnf instead of rpm because dnf does all the hard work of identifying all the dependencies and downloading anything else that’s needed, and importantly, validating the signatures and performing a test transaction before installing.
Neither of these databases are cleared when you run “dnf clean all”. That operation only removes all metadata and downloaded package files, things that can be re-downloaded on demand.
There are no “remote rpms” vs “local rpms”. You might say an RPM is installed, but the file used to install the package is not kept on the system, the files are extracted and placed on the filesystem, scripts are run, and the metadata is written to the database. After that, dnf deletes the downloaded package files from the cache, it is no longer needed.
Thanks Jonathan, until you mentioned that I hadn't thought about the implications of the download then update functionality that DNF does from the remote repositories. I guess that the two processes would use the same code path internally because there is no difference between the downloaded rpm and a pre-existing local rpm being installed, except for the deletion of the downloaded rpm, the local rpm doesn't get deleted.
regards, Steve
On Wed, 2024-11-27 at 08:30 +1100, Stephen Morris wrote:
My main concern was, if I use dnf to install a local rpm and afterwards I issue sudo dnf clean all to reclaim space (not that at the moment I have a need to be concerned about disk space) does that remove knowledge of the rpm having been installed?
No, the RPM database is separate.
Generally speaking, doing "clean all" is unnecessary for the user, and needlessly increases the workload on the repo servers. Now, instead of a user doing a metadata check and carrying on with perfectly fine cached data, it's fetching it all again. Multiply that by lots of users doing the same thing, and it's a huge increase. It gets even worse if people start downloading lots of RPMs, stuff something up, then dump their cache and start over again.
It was always bad advice, and became a bit of a cargo cult mentality.
When you do a dnf/yum update/install your software automatically handles caching of metadata and downloaded packages. It updates the metadata *when* needed, and this means out with the old in with the new, it's not accumulating more and more. Also, by default, the RPM packages it downloads to do these updates and installs are only kept on your drive until the job is finished, then they're purged.
On 11/24/24 2:04 PM, Stephen Morris wrote:
I have to questions around what that command is doing. 1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
If it hasn't been long enough since the last time it checked the repos, then it just loads the cached data. You can override that with the refresh option to force it to check.
2). When it does list repositories between the two messages, what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
If you don't use sudo, then it saves the cached data somewhere that your user has access to. I don't know where that is, but why would you want to do that anyway?
On 25/11/24 18:42, Samuel Sieb wrote:
On 11/24/24 2:04 PM, Stephen Morris wrote:
I have to questions around what that command is doing. 1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
If it hasn't been long enough since the last time it checked the repos, then it just loads the cached data. You can override that with the refresh option to force it to check.
2). When it does list repositories between the two messages, what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
If you don't use sudo, then it saves the cached data somewhere that your user has access to. I don't know where that is, but why would you want to do that anyway?
I'm not explicitly doing it, it's just that sometimes I forget to use sudo and dnf then goes off and does whatever it does for this command instead of prompting that it needs sudo to function, like some other software does.
regards, Steve
On 11/25/24 1:49 PM, Stephen Morris wrote:
On 25/11/24 18:42, Samuel Sieb wrote:
On 11/24/24 2:04 PM, Stephen Morris wrote:
I have to questions around what that command is doing. 1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
If it hasn't been long enough since the last time it checked the repos, then it just loads the cached data. You can override that with the refresh option to force it to check.
2). When it does list repositories between the two messages, what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
If you don't use sudo, then it saves the cached data somewhere that your user has access to. I don't know where that is, but why would you want to do that anyway?
I'm not explicitly doing it, it's just that sometimes I forget to use sudo and dnf then goes off and does whatever it does for this command instead of prompting that it needs sudo to function, like some other software does.
It doesn't need root access until it's affecting the system state. As a user, you can query the repos and download packages, but no install or remove.
On 26/11/24 11:32, Samuel Sieb wrote:
On 11/25/24 1:49 PM, Stephen Morris wrote:
On 25/11/24 18:42, Samuel Sieb wrote:
On 11/24/24 2:04 PM, Stephen Morris wrote:
I have to questions around what that command is doing. 1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
If it hasn't been long enough since the last time it checked the repos, then it just loads the cached data. You can override that with the refresh option to force it to check.
2). When it does list repositories between the two messages, what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
If you don't use sudo, then it saves the cached data somewhere that your user has access to. I don't know where that is, but why would you want to do that anyway?
I'm not explicitly doing it, it's just that sometimes I forget to use sudo and dnf then goes off and does whatever it does for this command instead of prompting that it needs sudo to function, like some other software does.
It doesn't need root access until it's affecting the system state. As a user, you can query the repos and download packages, but no install or remove.
My assumption was that when it said it was refreshing and loading the repositories that it was doing that in the system state, it wasn't until I queried what it was doing that I got the info that there is a current user interface to that functionality. Which then begs the question, if I issue the command "sudo dnf clean all" which cleans out the system state, does it also clean out the current user's data if it exists as well, or do you have to issue the command not under sudo as well for that to happen?
regards, Steve
On Wed, 2024-11-27 at 08:38 +1100, Stephen Morris wrote:
Which then begs the question, if I issue the command "sudo dnf clean all" which cleans out the system state, does it also clean out the current user's data if it exists as well, or do you have to issue the command not under sudo as well for that to happen?
There's an old joke: Patient: Doctor, doctor, it hurts when I do this. Doctor: Well, stop doing that, then.
But serious, try out the command both ways, then stop doing it if it worries you where it's storing the info.
*Very* old info suggested:
/var/cache/dnf is when dnf is used with superuser /var/tmp/dnf-<your_user_name>-<hash> is when you run it as you
Have a look, see if that's still the case.
On 27/11/24 13:08, Tim via users wrote:
On Wed, 2024-11-27 at 08:38 +1100, Stephen Morris wrote:
Which then begs the question, if I issue the command "sudo dnf clean all" which cleans out the system state, does it also clean out the current user's data if it exists as well, or do you have to issue the command not under sudo as well for that to happen?
There's an old joke: Patient: Doctor, doctor, it hurts when I do this. Doctor: Well, stop doing that, then.
But serious, try out the command both ways, then stop doing it if it worries you where it's storing the info.
*Very* old info suggested:
/var/cache/dnf is when dnf is used with superuser /var/tmp/dnf-<your_user_name>-<hash> is when you run it as youHave a look, see if that's still the case.
I did check this out and as you have suggested there is data in both sets of paths but the data in /var/tmp/dnf-<your_user_name>-<hash> is not maintained by the current dnf package. Looking at /var/cache/dnf begs the question of where is it specified for each repository how many cached copies dnf is going to keep. For the Fedora, Updates, Fedora Cisco, Rpmfusion-Free and Rpmfusion-Nonfree it seems to be keeping 5 copies, but for non-Fedora repos it seems to only be keeping 1 copy?
regards, Steve
On 11/27/24 1:29 PM, Stephen Morris wrote:
On 27/11/24 13:08, Tim via users wrote:
*Very* old info suggested:
/var/cache/dnf is when dnf is used with superuser /var/tmp/dnf-<your_user_name>-<hash> is when you run it as youHave a look, see if that's still the case.
I did check this out and as you have suggested there is data in both sets of paths but the data in /var/tmp/dnf-<your_user_name>-<hash> is not maintained by the current dnf package. Looking at /var/cache/dnf begs the question of where is it specified for each repository how many cached copies dnf is going to keep. For the Fedora, Updates, Fedora Cisco, Rpmfusion-Free and Rpmfusion-Nonfree it seems to be keeping 5 copies, but for non-Fedora repos it seems to only be keeping 1 copy?
It looks like it has separate directories for different mirrors. The non-Fedora repos probably only have a single server.
Stephen Morris:
Looking at /var/cache/dnf begs the question of where is it specified for each repository how many cached copies dnf is going to keep. For the Fedora, Updates, Fedora Cisco, Rpmfusion-Free and Rpmfusion-Nonfree it seems to be keeping 5 copies, but for non-Fedora repos it seems to only be keeping 1 copy?
Samuel Sieb:
It looks like it has separate directories for different mirrors. The non-Fedora repos probably only have a single server.
That makes sense, each mirror may not be identical to each other at any given moment in time. I would imagine that DNF manages each set of cached data automatically whenever it refers back to the same mirror.
On 11/26/24 1:38 PM, Stephen Morris wrote:
On 26/11/24 11:32, Samuel Sieb wrote:
On 11/25/24 1:49 PM, Stephen Morris wrote:
On 25/11/24 18:42, Samuel Sieb wrote:
On 11/24/24 2:04 PM, Stephen Morris wrote:
I have to questions around what that command is doing. 1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
If it hasn't been long enough since the last time it checked the repos, then it just loads the cached data. You can override that with the refresh option to force it to check.
2). When it does list repositories between the two messages, what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
If you don't use sudo, then it saves the cached data somewhere that your user has access to. I don't know where that is, but why would you want to do that anyway?
I'm not explicitly doing it, it's just that sometimes I forget to use sudo and dnf then goes off and does whatever it does for this command instead of prompting that it needs sudo to function, like some other software does.
It doesn't need root access until it's affecting the system state. As a user, you can query the repos and download packages, but no install or remove.
My assumption was that when it said it was refreshing and loading the repositories that it was doing that in the system state, it wasn't until I queried what it was doing that I got the info that there is a current user interface to that functionality. Which then begs the question, if I issue the command "sudo dnf clean all" which cleans out the system state, does it also clean out the current user's data if it exists as well, or do you have to issue the command not under sudo as well for that to happen?
If you run it under sudo, it has no knowledge of a user running it. You will have to do the clean as the user if you want to do that. Apparently I did test this at some point because there's a dnf-* directory owned by my user in /var/tmp.
On 27/11/24 14:19, Samuel Sieb wrote:
On 11/26/24 1:38 PM, Stephen Morris wrote:
On 26/11/24 11:32, Samuel Sieb wrote:
On 11/25/24 1:49 PM, Stephen Morris wrote:
On 25/11/24 18:42, Samuel Sieb wrote:
On 11/24/24 2:04 PM, Stephen Morris wrote:
I have to questions around what that command is doing. 1). Why is it producing the messages "Updating and loading repositories:"and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
If it hasn't been long enough since the last time it checked the repos, then it just loads the cached data. You can override that with the refresh option to force it to check.
2). When it does list repositories between the two messages, what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
If you don't use sudo, then it saves the cached data somewhere that your user has access to. I don't know where that is, but why would you want to do that anyway?
I'm not explicitly doing it, it's just that sometimes I forget to use sudo and dnf then goes off and does whatever it does for this command instead of prompting that it needs sudo to function, like some other software does.
It doesn't need root access until it's affecting the system state. As a user, you can query the repos and download packages, but no install or remove.
My assumption was that when it said it was refreshing and loading the repositories that it was doing that in the system state, it wasn't until I queried what it was doing that I got the info that there is a current user interface to that functionality. Which then begs the question, if I issue the command "sudo dnf clean all" which cleans out the system state, does it also clean out the current user's data if it exists as well, or do you have to issue the command not under sudo as well for that to happen?
If you run it under sudo, it has no knowledge of a user running it. You will have to do the clean as the user if you want to do that. Apparently I did test this at some point because there's a dnf-* directory owned by my user in /var/tmp.
I checked this out as well, there is a dnf-* directory owned by me in /var/tmp but it is not maintained by the current dnf. Issuing dnf clean all, not under sudo, removed 142 files, 75 directories with 0 errors, but it did not remove the directory in /var/tmp. This folder may not be a folder created by dnf, as the format is dnf -<my user name>-<entry> where "entry" doesn't look to be a hash, as for me "entry" is "erpyrkri".
regards, Steve
On 11/27/24 1:35 PM, Stephen Morris wrote:
On 27/11/24 14:19, Samuel Sieb wrote:
If you run it under sudo, it has no knowledge of a user running it. You will have to do the clean as the user if you want to do that. Apparently I did test this at some point because there's a dnf-* directory owned by my user in /var/tmp.
I checked this out as well, there is a dnf-* directory owned by me in / var/tmp but it is not maintained by the current dnf. Issuing dnf clean all, not under sudo, removed 142 files, 75 directories with 0 errors, but it did not remove the directory in /var/tmp. This folder may not be a folder created by dnf, as the format is dnf - <my user name>-<entry> where "entry" doesn't look to be a hash, as for me "entry" is "erpyrkri".
It won't remove the top-level directory, just like it won't remove /var/cache/dnf when you clean as root.
On 28/11/24 20:08, Samuel Sieb wrote:
On 11/27/24 1:35 PM, Stephen Morris wrote:
On 27/11/24 14:19, Samuel Sieb wrote:
If you run it under sudo, it has no knowledge of a user running it. You will have to do the clean as the user if you want to do that. Apparently I did test this at some point because there's a dnf-* directory owned by my user in /var/tmp.
I checked this out as well, there is a dnf-* directory owned by me in / var/tmp but it is not maintained by the current dnf. Issuing dnf clean all, not under sudo, removed 142 files, 75 directories with 0 errors, but it did not remove the directory in /var/tmp. This folder may not be a folder created by dnf, as the format is dnf
- <my user name>-<entry> where "entry" doesn't look to be a hash, as
for me "entry" is "erpyrkri".
It won't remove the top-level directory, just like it won't remove /var/cache/dnf when you clean as root.
I've just checked this folder and the dnf clean all didn't remove the contents of the folder, and it looks to be info for all the repositories I have. It may have been created when I had the python3-dnf-plugins-local package installed as one of the sub-folders is prefix by _dnf_local_, hence I'm surprised dnf clean all didn't remove the contents, although dnf5 doesn't put is cache in that folder so what I'm seeing may not be a surprise to other people, hence the /var/tmp folder may have been created by dnf4. I can delete the folder manually, but is this sort of thing normal when upgrading from say F40 to F41?
regards, Steve
On Mon, 2024-11-25 at 09:04 +1100, Stephen Morris wrote:
1). Why is it producing the messages "Updating and loadingrepositories:" and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
There are probably various things it can do that don't produce any individual progress status. Just a did this, did that, and this...
When you fire up dnf, it is going to load data, some of it will be local, some of it may be fetched from the WWW.
If you want to look for differences. Purge your dnf cache, do a dnf update (check for updates), watch what happens. Say no and abort. Do another dnf update moments later, and see how it operates from cached data (what progress info you see as it does its thing).
2). When it does list repositories between the two messages,what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
If you do a dnf <anything> as a normal user (e.g. look to see if some package exists, or info about it, or if there's any updates, etc), it'll have to get data from the remote server because the info it previously got as a root user isn't available to it.
If you then do it as root (whether sudo, or su), it's going to have to get the metadata again, because the data it fetched and cached as an ordinary user isn't available to it.
Yes, I've occasionally done a dnf search httpd (for example) as an ordinary user, because I couldn't be bothered typing in a password with sudo. But then, if I do want to download and install something, I'm going to have to do that, anyway. So there's little advantage in it.
On 25/11/24 22:43, Tim wrote:
On Mon, 2024-11-25 at 09:04 +1100, Stephen Morris wrote:
1). Why is it producing the messages "Updating and loadingrepositories:" and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
There are probably various things it can do that don't produce any individual progress status. Just a did this, did that, and this...
When you fire up dnf, it is going to load data, some of it will be local, some of it may be fetched from the WWW.
If you want to look for differences. Purge your dnf cache, do a dnf update (check for updates), watch what happens. Say no and abort. Do another dnf update moments later, and see how it operates from cached data (what progress info you see as it does its thing).
I'm just issuing this command to see if there are actually any updates to be applied rather than having dnf upgrade tell me there is nothing to do.
2). When it does list repositories between the two messages,what is it actually doing when the command is not issued under sudo, particularly when issuing the command lists the two updates repositories and then issuing it again immediately after produces nothing between the two messages?
If you do a dnf <anything> as a normal user (e.g. look to see if some package exists, or info about it, or if there's any updates, etc), it'll have to get data from the remote server because the info it previously got as a root user isn't available to it.
If you then do it as root (whether sudo, or su), it's going to have to get the metadata again, because the data it fetched and cached as an ordinary user isn't available to it.
Yes, I've occasionally done a dnf search httpd (for example) as an ordinary user, because I couldn't be bothered typing in a password with sudo. But then, if I do want to download and install something, I'm going to have to do that, anyway. So there's little advantage in it.
I assumed that when dnf retrieved it's metadata it was updating the standard system locations which a normal user can't update, I didn't know it was caching the data elsewhere. I'll need to do some more checking tomorrow, but I thought I have been in the situation where dnf check-upgrade loaded metadata and then sudo dnf check-upgrade said there was nothing to do.
regards, Steve
On 11/25/24 2:02 PM, Stephen Morris wrote:
On 25/11/24 22:43, Tim wrote:
On Mon, 2024-11-25 at 09:04 +1100, Stephen Morris wrote:
1). Why is it producing the messages "Updating and loadingrepositories:" and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
There are probably various things it can do that don't produce any individual progress status. Just a did this, did that, and this...
When you fire up dnf, it is going to load data, some of it will be local, some of it may be fetched from the WWW.
If you want to look for differences. Purge your dnf cache, do a dnf update (check for updates), watch what happens. Say no and abort. Do another dnf update moments later, and see how it operates from cached data (what progress info you see as it does its thing).
I'm just issuing this command to see if there are actually any updates to be applied rather than having dnf upgrade tell me there is nothing to do.
What's the benefit of that? If there is something, then you'll just have to run another command anyway.
On 26/11/24 11:35, Samuel Sieb wrote:
On 11/25/24 2:02 PM, Stephen Morris wrote:
On 25/11/24 22:43, Tim wrote:
On Mon, 2024-11-25 at 09:04 +1100, Stephen Morris wrote:
1). Why is it producing the messages "Updating and loading repositories:" and "Repositories loaded." with nothing between them which means that it didn't do anything? This functionality is independent of whether the command is issued under sudo or not.
There are probably various things it can do that don't produce any individual progress status. Just a did this, did that, and this...
When you fire up dnf, it is going to load data, some of it will be local, some of it may be fetched from the WWW.
If you want to look for differences. Purge your dnf cache, do a dnf update (check for updates), watch what happens. Say no and abort. Do another dnf update moments later, and see how it operates from cached data (what progress info you see as it does its thing).
I'm just issuing this command to see if there are actually any updates to be applied rather than having dnf upgrade tell me there is nothing to do.
What's the benefit of that? If there is something, then you'll just have to run another command anyway.
This is true, and I know this can be done under DNF, but I'm using that to decide whether it is worth doing the upgrade in terms of the volume of updates that will be put on. If there are only 3 updates to put on then its not necessarily worth putting on the updates yet.
regards, Steve
On Wed, 2024-11-27 at 08:46 +1100, Stephen Morris wrote:
This is true, and I know this can be done under DNF, but I'm using that to decide whether it is worth doing the upgrade in terms of the volume of updates that will be put on. If there are only 3 updates to put on then its not necessarily worth putting on the updates yet.
So, just do "sudo dnf update" and when it comes back with the results, look at them, THEN hit Y or N for yes or no to install them.
People who preload the install/update/erase command line with yes are just asking for trouble. If something goes rogue, a command may install or remove hundreds of packages, leaving you with a borked system.
That's another bit of bad advice various websites offer.
On 27/11/24 12:48, Tim via users wrote:
On Wed, 2024-11-27 at 08:46 +1100, Stephen Morris wrote:
This is true, and I know this can be done under DNF, but I'm using that to decide whether it is worth doing the upgrade in terms of the volume of updates that will be put on. If there are only 3 updates to put on then its not necessarily worth putting on the updates yet.
So, just do "sudo dnf update" and when it comes back with the results, look at them, THEN hit Y or N for yes or no to install them.
People who preload the install/update/erase command line with yes are just asking for trouble. If something goes rogue, a command may install or remove hundreds of packages, leaving you with a borked system.
That's another bit of bad advice various websites offer.
I never use -y. I know I could issue dnf upgrade and do the same thing, I've just been playing around with dnf check-upgrade. I could be wrong but dnf check-upgrade seems to me to be new with dnf5, so presumably it has been added for a reason, so is there checking done that is not done with dnf upgrade?
Just an off topic question, I've noticed recently that when I preview a mail from this list I no longer get any indication of who the email came from in the body of the mail, and if there is no signature, with the from being "Community support for Fedora users", it makes it hard to identify who we are communicating with. Having said this though, when I reply to the mail, I do get an indication of who the mail came from that I used to get in the preview before replying. Has this been a change in functionality of the mail list?
regards, Steve
On Wed, Nov 27, 2024 at 4:57 PM Stephen Morris steve.morris.au@gmail.com wrote:
[...] dnf check-upgrade seems to me to be new with dnf5
I've been using dnf check-upgrade for years.
On 28/11/24 09:15, Go Canes wrote:
On Wed, Nov 27, 2024 at 4:57 PM Stephen Morris steve.morris.au@gmail.com wrote:
[...] dnf check-upgrade seems to me to be new with dnf5
I've been using dnf check-upgrade for years.
Thankyou. I'd only noticed that command as a result of issuing dnf --help since I upgraded to F41.
regards, Steve
On Thu, 2024-11-28 at 08:57 +1100, Stephen Morris wrote:
Just an off topic question, I've noticed recently that when I preview a mail from this list I no longer get any indication of who the email came from in the body of the mail, and if there is no signature, with the from being "Community support for Fedora users", it makes it hard to identify who we are communicating with. Having said this though, when I reply to the mail, I do get an indication of who the mail came from that I used to get in the preview before replying. Has this been a change in functionality of the mail list?
Nothing has changed on the list that I can see, other than how some people are using it (not trimming quotes of quotes in replies when they should be). My guess would be it's your mail program suppressing or hiding quoted text, some have options for that.
When I quote someone I have their name directly above their text. Making it very easy to slice of any older quotes of quotes without losing the author's name. I usually avoid quotes of quotes, perhaps excepting for brief snippets where it helps explain what's written after it. Again keeping their name directly above their content. So the whole thing's easier to cut out without lots of editing here, there, and everywhere.
In the *headers* of the mail some author's email addresses do get removed by the mailing list (but not their name). It's to do with posters sending from addresses that cause problems with some anti-spam systems (since the distributed list mail will actually be coming from the list, not directly from the author).
Stephen Morris wrote:
Just an off topic question, I've noticed recently that when I preview a mail from this list I no longer get any indication of who the email came from in the body of the mail, and if there is no signature, with the from being "Community support for Fedora users"
Tim :
Nothing has changed on the list that I can see, other than how some people are using it (not trimming quotes of quotes in replies when they should be). My guess would be it's your mail program suppressing or hiding quoted text, some have options for that.
Also, look for options regarding suppressing/hiding signatures (that's text after a dash dash blank space new-line sequence, the traditional signature separator character sequence).
On 28/11/24 20:54, Tim via users wrote:
Stephen Morris wrote:
Just an off topic question, I've noticed recently that when I preview a mail from this list I no longer get any indication of who the email came from in the body of the mail, and if there is no signature, with the from being "Community support for Fedora users"
Tim :
Nothing has changed on the list that I can see, other than how some people are using it (not trimming quotes of quotes in replies when they should be). My guess would be it's your mail program suppressing or hiding quoted text, some have options for that.
Also, look for options regarding suppressing/hiding signatures (that's text after a dash dash blank space new-line sequence, the traditional signature separator character sequence).
What I was looking for as a signature was something like I specify at the bottom of my mails, which seems to be missing from a lot of other people's mails. Looking at your reply, what I see in your reply is completely different to what I see in Samuel's reply below, and in this reply the first line that shows the last replier to this thread was you, that line does not appear in the mail preview nor if I double click the mail to load it into a separate tab, it only shows when I actually reply to the mail. The mail header that Thunderbird shows has you in the CC which is potentially an indication the mail came from you, but that is no guarantee.
regards, Steve
On 11/29/24 4:40 PM, Stephen Morris wrote:
On 28/11/24 20:54, Tim via users wrote:
Stephen Morris wrote:
Just an off topic question, I've noticed recently that when I preview a mail from this list I no longer get any indication of who the email came from in the body of the mail, and if there is no signature, with the from being "Community support for Fedora users"
Tim :
Nothing has changed on the list that I can see, other than how some people are using it (not trimming quotes of quotes in replies when they should be). My guess would be it's your mail program suppressing or hiding quoted text, some have options for that.
Also, look for options regarding suppressing/hiding signatures (that's text after a dash dash blank space new-line sequence, the traditional signature separator character sequence).
What I was looking for as a signature was something like I specify at the bottom of my mails, which seems to be missing from a lot of other people's mails. Looking at your reply, what I see in your reply is completely different to what I see in Samuel's reply below, and in this reply the first line that shows the last replier to this thread was you, that line does not appear in the mail preview nor if I double click the mail to load it into a separate tab, it only shows when I actually reply to the mail. The mail header that Thunderbird shows has you in the CC which is potentially an indication the mail came from you, but that is no guarantee.
This was explained in another email. Certain mail hosts cause issues with mailing lists, so the list changes the from to be from the list and puts the sending user in the CC. And when you reply, you see the name "via users".
On 30/11/24 13:27, Samuel Sieb wrote:
On 11/29/24 4:40 PM, Stephen Morris wrote:
On 28/11/24 20:54, Tim via users wrote:
Stephen Morris wrote:
Just an off topic question, I've noticed recently that when I preview a mail from this list I no longer get any indication of who the email came from in the body of the mail, and if there is no signature, with the from being "Community support for Fedora users"
Tim :
Nothing has changed on the list that I can see, other than how some people are using it (not trimming quotes of quotes in replies when they should be). My guess would be it's your mail program suppressing or hiding quoted text, some have options for that.
Also, look for options regarding suppressing/hiding signatures (that's text after a dash dash blank space new-line sequence, the traditional signature separator character sequence).
What I was looking for as a signature was something like I specify at the bottom of my mails, which seems to be missing from a lot of other people's mails. Looking at your reply, what I see in your reply is completely different to what I see in Samuel's reply below, and in this reply the first line that shows the last replier to this thread was you, that line does not appear in the mail preview nor if I double click the mail to load it into a separate tab, it only shows when I actually reply to the mail. The mail header that Thunderbird shows has you in the CC which is potentially an indication the mail came from you, but that is no guarantee.
This was explained in another email. Certain mail hosts cause issues with mailing lists, so the list changes the from to be from the list and puts the sending user in the CC. And when you reply, you see the name "via users".
Thankyou. I didn't realise the name in the CC was supplied by the list. What I'm seeing in the body of the mail, I believe, is the html version of the mail, not the text version, unless Thunderbird is manipulating the incoming mail to colour code the response indicators, but the body is behaving differently to what I remember although I could be remembering things incorrectly.
regards, Steve
On Sat, 2024-11-30 at 11:40 +1100, Stephen Morris wrote:
What I was looking for as a signature was something like I specify at the bottom of my mails, which seems to be missing from a lot of other people's mails.
Ah! PGP... Few do that (that I see).
For the most part, it's a waste of time. Unless you've collected a signature *directly* from me, there's little confidence that someone saying that they're me, even if signed, is me.
There's a tiny bit of confidence when someone else has vouched for me, and countersigned. But how many of them really check?
There's virtually no confidence in signatures from PGP servers, as far as I can see, unless they're countersigned by various people that I would know or trust. And I really don't know anyone on this list. I've never met any of them. And if you did, could you tell a real ID from a fake ID if they tried to prove they were who they said they were?
I also found that PGP servers were a major source of spam. Any time you uploaded a key, somebody harvested it and spammed me, straight away. I could test this by creating extra keys and uploading them, they'd get harvested and spammed in minutes.
Looking at your reply, what I see in your reply is completely different to what I see in Samuel's reply below, and in this reply the first line that shows the last replier to this thread was you, that line does not appear in the mail preview nor if I double click the mail to load it into a separate tab, it only shows when I actually reply to the mail.
If you're talking about body text, everyone else has their style. Some have a stack of intro lines like:
last person said:
second to last person wrote:
third to last person posted:
Which gets messy real quick. I try to limit replies to quoting no more than two people if it makes the message more coherent than just one person's last comments. And try to allow newspaper style editing - you can slice off everything above and below the bits you're concerned with, in huge slabs, you don't have to hunt and peck through.
And again, I think you may be seeing something like your mail program hiding quoted text, even signatures (the textual boilerplates, not PGP signing). If I look on the mail archive, it's definitely doing that (which probably explains why a lot of people put zero effort into trimming their replies). If you click on the (...) on the webpage you can expand the hidden text, but that makes for laborious reading of messages.
As a comparison, have a look through your menus for "view source" and see what a message looks like before Thunderbird converts it into its own HTML rendition of the message (regardless of whether it was HTML or plain text).
Thunderbirds re-rendering of messages is yet another reason that I don't like it.
The mail header that Thunderbird shows has you in the CC which is potentially an indication the mail came from you, but that is no guarantee.
That's already been explained that a few times.
To be honest, you can't trust any "from" address as being real. Anyone can fake the headers, few things verify them, and there's still plenty of mail servers that don't verify that you are who you say you are before accepting a post. The SMTP servers that I use do require my mail program to log-in when posting, and my from address has to match, but there's more to it than just that.
On 30/11/24 20:05, Tim via users wrote:
On Sat, 2024-11-30 at 11:40 +1100, Stephen Morris wrote:
What I was looking for as a signature was something like I specify at the bottom of my mails, which seems to be missing from a lot of other people's mails.
Ah! PGP... Few do that (that I see).
For the most part, it's a waste of time. Unless you've collected a signature *directly* from me, there's little confidence that someone saying that they're me, even if signed, is me.
There's a tiny bit of confidence when someone else has vouched for me, and countersigned. But how many of them really check?
There's virtually no confidence in signatures from PGP servers, as far as I can see, unless they're countersigned by various people that I would know or trust. And I really don't know anyone on this list. I've never met any of them. And if you did, could you tell a real ID from a fake ID if they tried to prove they were who they said they were?
I also found that PGP servers were a major source of spam. Any time you uploaded a key, somebody harvested it and spammed me, straight away. I could test this by creating extra keys and uploading them, they'd get harvested and spammed in minutes.
Looking at your reply, what I see in your reply is completely different to what I see in Samuel's reply below, and in this reply the first line that shows the last replier to this thread was you, that line does not appear in the mail preview nor if I double click the mail to load it into a separate tab, it only shows when I actually reply to the mail.
If you're talking about body text, everyone else has their style. Some have a stack of intro lines like:
last person said:
second to last person wrote:
third to last person posted:
Which gets messy real quick. I try to limit replies to quoting no more than two people if it makes the message more coherent than just one person's last comments. And try to allow newspaper style editing - you can slice off everything above and below the bits you're concerned with, in huge slabs, you don't have to hunt and peck through.
The issue with this is what you have shown is what Thunderbird used to always show to identify who said what which I find extremely useful to understand response history and the context of replies. But recently I've seen two different forms of that from this list, so I don't know whether that was list changes, a daily version change in Thunderbird, or whether the respondent changed the format of the response histories by snipping specific portions.
And again, I think you may be seeing something like your mail program hiding quoted text, even signatures (the textual boilerplates, not PGP signing). If I look on the mail archive, it's definitely doing that (which probably explains why a lot of people put zero effort into trimming their replies). If you click on the (...) on the webpage you can expand the hidden text, but that makes for laborious reading of messages.
As a comparison, have a look through your menus for "view source" and see what a message looks like before Thunderbird converts it into its own HTML rendition of the message (regardless of whether it was HTML or plain text).
Thunderbirds re-rendering of messages is yet another reason that I don't like it.
I did a "view source" and after trawling through all the headers to get to what looks like your response there is no plain text and no html, it displays like encrypted data, there is nothing visible at all the matches your text.
The mail header that Thunderbird shows has you in the CC which is potentially an indication the mail came from you, but that is no guarantee.
That's already been explained that a few times.
To be honest, you can't trust any "from" address as being real. Anyone can fake the headers, few things verify them, and there's still plenty of mail servers that don't verify that you are who you say you are before accepting a post. The SMTP servers that I use do require my mail program to log-in when posting, and my from address has to match, but there's more to it than just that.
Yes, I understand that, I mask the from address all the time in mails sent out by programs I have written at work, so that there is a meaningful address for recipients to reply to rather than the sudo mail address of the scheduler the programs are running under.
regards, Steve
On 11/30/24 3:04 PM, Stephen Morris wrote:
On 30/11/24 20:05, Tim via users wrote:
As a comparison, have a look through your menus for "view source" and see what a message looks like before Thunderbird converts it into its own HTML rendition of the message (regardless of whether it was HTML or plain text).
Thunderbirds re-rendering of messages is yet another reason that I don't like it.
I did a "view source" and after trawling through all the headers to get to what looks like your response there is no plain text and no html, it displays like encrypted data, there is nothing visible at all the matches your text.
Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64
The plain text is encoded as base64.
On 1/12/24 10:32, Samuel Sieb wrote:
On 11/30/24 3:04 PM, Stephen Morris wrote:
On 30/11/24 20:05, Tim via users wrote:
As a comparison, have a look through your menus for "view source" and see what a message looks like before Thunderbird converts it into its own HTML rendition of the message (regardless of whether it was HTML or plain text).
Thunderbirds re-rendering of messages is yet another reason that I don't like it.
I did a "view source" and after trawling through all the headers to get to what looks like your response there is no plain text and no html, it displays like encrypted data, there is nothing visible at all the matches your text.
Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64
The plain text is encoded as base64.
Thanks Samuel, I understood that the mails were mime encoded, but when I viewed the source to see whether when I previewed or opened the mail I was looking at plain text or a html version which is my preferred version, I was expecting to see the body of the email as either plain text or html tags, particularly when viewing the mail makes it look like the client is working with html.
regards, Steve
On 11/30/24 3:44 PM, Stephen Morris wrote:
On 1/12/24 10:32, Samuel Sieb wrote:
On 11/30/24 3:04 PM, Stephen Morris wrote:
On 30/11/24 20:05, Tim via users wrote:
As a comparison, have a look through your menus for "view source" and see what a message looks like before Thunderbird converts it into its own HTML rendition of the message (regardless of whether it was HTML or plain text).
Thunderbirds re-rendering of messages is yet another reason that I don't like it.
I did a "view source" and after trawling through all the headers to get to what looks like your response there is no plain text and no html, it displays like encrypted data, there is nothing visible at all the matches your text.
Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64
The plain text is encoded as base64.
Thanks Samuel, I understood that the mails were mime encoded, but when I viewed the source to see whether when I previewed or opened the mail I was looking at plain text or a html version which is my preferred version, I was expecting to see the body of the email as either plain text or html tags, particularly when viewing the mail makes it look like the client is working with html.
The funny thing with your emails is that the plain text part is base64 encoded and the html part is not.
On 1/12/24 10:53, Samuel Sieb wrote:
On 11/30/24 3:44 PM, Stephen Morris wrote:
On 1/12/24 10:32, Samuel Sieb wrote:
On 11/30/24 3:04 PM, Stephen Morris wrote:
On 30/11/24 20:05, Tim via users wrote:
As a comparison, have a look through your menus for "view source" and see what a message looks like before Thunderbird converts it into its own HTML rendition of the message (regardless of whether it was HTML or plain text).
Thunderbirds re-rendering of messages is yet another reason that I don't like it.
I did a "view source" and after trawling through all the headers to get to what looks like your response there is no plain text and no html, it displays like encrypted data, there is nothing visible at all the matches your text.
Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64
The plain text is encoded as base64.
Thanks Samuel, I understood that the mails were mime encoded, but when I viewed the source to see whether when I previewed or opened the mail I was looking at plain text or a html version which is my preferred version, I was expecting to see the body of the email as either plain text or html tags, particularly when viewing the mail makes it look like the client is working with html.
The funny thing with your emails is that the plain text part is base64 encoded and the html part is not.
That's interesting, I would have expected both to be the same. I used to use the "Auto" compose option, which supposedly sends the mail as plain text if there is no formatting otherwise it sends it as html, but after I think you requested that I stop sending through html, I changed the option to send both html and text for the relevant client to decide which one it was going to use, hence I would have expected both to be encoded, which I thought was a mail standard. I'm using a version of Thunderbird that is several versions ahead of what's in the Fedora repositories, and I think I'm on the mail list for Thunderbird, so I'll raise it on there to see if I get any feedback.
regards, Steve
On 11/30/24 4:24 PM, Stephen Morris wrote:
On 1/12/24 10:53, Samuel Sieb wrote:
On 11/30/24 3:44 PM, Stephen Morris wrote:
On 1/12/24 10:32, Samuel Sieb wrote:
On 11/30/24 3:04 PM, Stephen Morris wrote:
On 30/11/24 20:05, Tim via users wrote:
As a comparison, have a look through your menus for "view source" and see what a message looks like before Thunderbird converts it into its own HTML rendition of the message (regardless of whether it was HTML or plain text).
Thunderbirds re-rendering of messages is yet another reason that I don't like it.
I did a "view source" and after trawling through all the headers to get to what looks like your response there is no plain text and no html, it displays like encrypted data, there is nothing visible at all the matches your text.
Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64
The plain text is encoded as base64.
Thanks Samuel, I understood that the mails were mime encoded, but when I viewed the source to see whether when I previewed or opened the mail I was looking at plain text or a html version which is my preferred version, I was expecting to see the body of the email as either plain text or html tags, particularly when viewing the mail makes it look like the client is working with html.
The funny thing with your emails is that the plain text part is base64 encoded and the html part is not.
That's interesting, I would have expected both to be the same. I used to use the "Auto" compose option, which supposedly sends the mail as plain text if there is no formatting otherwise it sends it as html, but after I think you requested that I stop sending through html, I changed the option to send both html and text for the relevant client to decide which one it was going to use, hence I would have expected both to be encoded, which I thought was a mail standard. I'm using a version of Thunderbird that is several versions ahead of what's in the Fedora repositories, and I think I'm on the mail list for Thunderbird, so I'll raise it on there to see if I get any feedback.
There's nothing wrong with it. base64 encoding is normal. I just thought it was funny that only the plain text was encoded like that.
On Sun, 2024-12-01 at 10:04 +1100, Stephen Morris wrote:
The issue with this is what you have shown is what Thunderbird used to always show to identify who said what which I find extremely useful to understand response history and the context of replies.
Again, look through your preferences, and see if there are options about hiding/muting/suppressing quoted text. Various mail programs have that option.
This email starts off with a line saying the date, time, and your name, quotes one paragraph from you before my reply, and quotes one more paragraph from you between the next one where I started with "Also," and the next I started with "My messages".
Also, leave a blank line between quotes and your replies. It is very hard to read mail when everything is just one huge block.
I did a "view source" and after trawling through all the headers to get to what looks like your response there is no plain text and no html, it displays like encrypted data, there is nothing visible at all the matches your text.
My messages are sent as plain text. And it was sent as 7-bit text. So, somewhere between me and you it has been converted.
Various mail servers try to be helpful doing that (turning things from 8-bit to 7-bit, vice versa, quoted-printable, base 64) when it's mostly not necessary, and sometimes cause problems, when they bugger it up. Whereas a human would have been able to cope with an odd character error here and there.
And I think Thunderbird creates a HTML render on the fly, but doesn't change the content as it came in.
Tim:
Also, leave a blank line between quotes and your replies. It is very hard to read mail when everything is just one huge block.
Patrick O'Callaghan:
This. Drives me crazy. How hard is it to hit a single Return before starting to type?
Twice is even better...
Quite apart from making a clear distinction between quotes and new content, for humans to read easier. Many mail programs will join the reply into being part of the quote without the blank line between them.
On 1/12/24 22:23, Tim via users wrote:
Tim:
Also, leave a blank line between quotes and your replies. It is very hard to read mail when everything is just one huge block.
Patrick O'Callaghan:
This. Drives me crazy. How hard is it to hit a single Return before starting to type?
Twice is even better...
Quite apart from making a clear distinction between quotes and new content, for humans to read easier. Many mail programs will join the reply into being part of the quote without the blank line between them.
Sorry, this is my fault. I'm used to working with html mails where the response is black text and the quotes are colour coded for quotes from different responders, and I keep forgetting that environments that insist on using text only potentially don't have the colour coding to make things obvious.
Just on the formatting front, the way the quotes are formatted, with Tim's name above his quote and Patrick's name above his quote, from what I'm seeing is a recent departure from what I'm used to seeing as standard for quotes.
regards, Steve
On Mon, 2024-12-02 at 08:54 +1100, Stephen Morris wrote:
Quite apart from making a clear distinction between quotes and new content, for humans to read easier. Many mail programs will join the reply into being part of the quote without the blank line between them.
Sorry, this is my fault. I'm used to working with html mails where the response is black text and the quotes are colour coded for quotes from different responders, and I keep forgetting that environments that insist on using text only potentially don't have the colour coding to make things obvious.
Just on the formatting front, the way the quotes are formatted, with Tim's name above his quote and Patrick's name above his quote, from what I'm seeing is a recent departure from what I'm used to seeing as standard for quotes.
Not sure what you mean. The style you mention is the default in Evolution and other mailers, and is what I've been using for over 20 years.
poc
On 2/12/24 09:10, Patrick O'Callaghan wrote:
On Mon, 2024-12-02 at 08:54 +1100, Stephen Morris wrote:
Quite apart from making a clear distinction between quotes and new content, for humans to read easier. Many mail programs will join the reply into being part of the quote without the blank line between them.
Sorry, this is my fault. I'm used to working with html mails where the response is black text and the quotes are colour coded for quotes from different responders, and I keep forgetting that environments that insist on using text only potentially don't have the colour coding to make things obvious.
Just on the formatting front, the way the quotes are formatted, with Tim's name above his quote and Patrick's name above his quote, from what I'm seeing is a recent departure from what I'm used to seeing as standard for quotes.
Not sure what you mean. The style you mention is the default in Evolution and other mailers, and is what I've been using for over 20 years.
In the previous mail I was referring to, where Tim's name appears above his portion as "Tim:" in black and your name above your portion as "Patrick O'Callaghan:" also in black, which makes it stand out from the colour coding of the quotes, has only been exhibiting itself that way recently. I've only been seeing mails on this mail list being displayed that way within the last week. In the past none of your emails have ever displayed that way, they all displayed like this one I'm responding to, which doesn't have that same text header. I'm not saying that it is a problem, I'm just commenting that from my perspective it is not usual.
regards, Steve
poc
On Mon, 2024-12-02 at 09:24 +1100, Stephen Morris wrote:
Not sure what you mean. The style you mention is the default in Evolution and other mailers, and is what I've been using for over 20 years.
In the previous mail I was referring to, where Tim's name appears above his portion as "Tim:" in black and your name above your portion as "Patrick O'Callaghan:" also in black, which makes it stand out from the colour coding of the quotes, has only been exhibiting itself that way recently. I've only been seeing mails on this mail list being displayed that way within the last week. In the past none of your emails have ever displayed that way, they all displayed like this one I'm responding to, which doesn't have that same text header. I'm not saying that it is a problem, I'm just commenting that from my perspective it is not usual.
I wouldn't know. I never use HTML mail on mailing lists since most of them frown on it, including this one (read the Guidelines). I read the text alternative if present, otherwise I generally ignore the message, and I *never* post in HTML.
poc
On Mon, 2024-12-02 at 08:54 +1100, Stephen Morris wrote:
Sorry, this is my fault. I'm used to working with html mails where the response is black text and the quotes are colour coded for quotes from different responders, and I keep forgetting that environments that insist on using text only potentially don't have the colour coding to make things obvious.
I am seeing your HTML mail, and while it does highlight the edge of different sections, one huge slab of text makes it difficult to find new comments in the middle. And when replying to such things, quotes often get joined together that shouldn't.
It doesn't matter whether you post HTML or plain text, it helps things an awful lot when there's a gap between things. Just the same as it helps when people use paragraphs. I know someone on another list who can write three pages of text as one paragraph. My eyes run out of breath reading them.
My mail program does colour quotes different from new content, but with plain text mail all quotes are grey, new text is black. And if a poster didn't leave a gap between quotes and replies, often the reply is coloured the same as the quote. With HTML mail, it relies on the message styling itself. Yours are all black text, with coloured stripes running down the side.
I'm used to mailing lists being read by people using a diverse range of email clients, that all behave in different ways, and are often read by people with bad eyesight. Making things easy to follow makes life easier for everyone.
Just on the formatting front, the way the quotes are formatted, with Tim's name above his quote and Patrick's name above his quote, from what I'm seeing is a recent departure from what I'm used to seeing as standard for quotes.
Not everybody quotes the same. I edit my replies. I try to make it easy to tell who said what, rather than make people count the indents, etc. The way things often get done, people often get misquoted (saying someone else said what they said) with them all stacked up the top separate from the content. I try to make it so you can slice off quotes of quotes in one fell swoop, and still have the name of the last contributor above their own text. And usually edit out the quotes of quotes of quotes (like we're supposed to do), if they really aren't needed to follow what's written in a message.
On 2/12/24 22:39, Tim via users wrote:
On Mon, 2024-12-02 at 08:54 +1100, Stephen Morris wrote:
Sorry, this is my fault. I'm used to working with html mails where the response is black text and the quotes are colour coded for quotes from different responders, and I keep forgetting that environments that insist on using text only potentially don't have the colour coding to make things obvious.
I am seeing your HTML mail, and while it does highlight the edge of different sections, one huge slab of text makes it difficult to find new comments in the middle. And when replying to such things, quotes often get joined together that shouldn't.
It doesn't matter whether you post HTML or plain text, it helps things an awful lot when there's a gap between things. Just the same as it helps when people use paragraphs. I know someone on another list who can write three pages of text as one paragraph. My eyes run out of breath reading them.
My mail program does colour quotes different from new content, but with plain text mail all quotes are grey, new text is black. And if a poster didn't leave a gap between quotes and replies, often the reply is coloured the same as the quote. With HTML mail, it relies on the message styling itself. Yours are all black text, with coloured stripes running down the side.
That's interesting. In Thunderbird I'm seeing things differently, depending on who is reply to threads I'm seeing the entire quote with a black line down the side and text in black, for other people I see a coloured stripe and the text in blue. For the quotes you are replying to I see a dark blue stripe and light blue text with your response in black. What is also interesting is that for to portion of your response I'm replying to, in this composition that quote has a dark blue stripe and black text.
I'm used to mailing lists being read by people using a diverse range of email clients, that all behave in different ways, and are often read by people with bad eyesight. Making things easy to follow makes life easier for everyone.
Just on the formatting front, the way the quotes are formatted, with Tim's name above his quote and Patrick's name above his quote, from what I'm seeing is a recent departure from what I'm used to seeing as standard for quotes.
Not everybody quotes the same. I edit my replies. I try to make it easy to tell who said what, rather than make people count the indents, etc. The way things often get done, people often get misquoted (saying someone else said what they said) with them all stacked up the top separate from the content. I try to make it so you can slice off quotes of quotes in one fell swoop, and still have the name of the last contributor above their own text. And usually edit out the quotes of quotes of quotes (like we're supposed to do), if they really aren't needed to follow what's written in a message.
In my response to Patrick on this formatting, I've only started seeing your mails and one of Patrick's carved up that way in the last week, and like I said to Patrick, I'm not saying it is a problem I'm just saying from what I'm used to seeing on this mail list its unusual. I'm also used to seeing who the respondent was in a message above each quote, but that seems to have been changed recently as well.
regards, Steve
On Tue, 2024-12-03 at 09:15 +1100, Stephen Morris wrote:
In my response to Patrick on this formatting, I've only started seeing your mails and one of Patrick's carved up that way in the last week, and like I said to Patrick, I'm not saying it is a problem I'm just saying from what I'm used to seeing on this mail list its unusual. I'm also used to seeing who the respondent was in a message above each quote, but that seems to have been changed recently as well.
The mailing list software doesn't manipulate the body text of messages and has no conception of what a quote even is, and for my part as I've said I haven't changed anything either, so if you think something has changed it is presumably at your end. I can't see any difference between recent list messages and those of 5 or 10 years ago (I keep a full archive).
poc
On 12/2/24 2:15 PM, Stephen Morris wrote:
On 2/12/24 22:39, Tim via users wrote:
On Mon, 2024-12-02 at 08:54 +1100, Stephen Morris wrote:
Sorry, this is my fault. I'm used to working with html mails where the response is black text and the quotes are colour coded for quotes from different responders, and I keep forgetting that environments that insist on using text only potentially don't have the colour coding to make things obvious.
I am seeing your HTML mail, and while it does highlight the edge of different sections, one huge slab of text makes it difficult to find new comments in the middle. And when replying to such things, quotes often get joined together that shouldn't.
It doesn't matter whether you post HTML or plain text, it helps things an awful lot when there's a gap between things. Just the same as it helps when people use paragraphs. I know someone on another list who can write three pages of text as one paragraph. My eyes run out of breath reading them.
My mail program does colour quotes different from new content, but with plain text mail all quotes are grey, new text is black. And if a poster didn't leave a gap between quotes and replies, often the reply is coloured the same as the quote. With HTML mail, it relies on the message styling itself. Yours are all black text, with coloured stripes running down the side.
That's interesting. In Thunderbird I'm seeing things differently, depending on who is reply to threads I'm seeing the entire quote with a black line down the side and text in black, for other people I see a coloured stripe and the text in blue. For the quotes you are replying to I see a dark blue stripe and light blue text with your response in black. What is also interesting is that for to portion of your response I'm replying to, in this composition that quote has a dark blue stripe and black text.
I think that's the difference between Thunderbird showing a plain text message vs. an HTML message. Since you send both, I can switch between which one is shown and see the difference as you describe. As someone else mentioned, with HTML messages Thunderbird can only show it as it was sent, so that's why you get black quoted text. With plain text, it identifies the quotes, colors the text differently, and creates the quote bars on the left with different colors.
On 3/12/24 09:55, Samuel Sieb wrote:
On 12/2/24 2:15 PM, Stephen Morris wrote:
On 2/12/24 22:39, Tim via users wrote:
On Mon, 2024-12-02 at 08:54 +1100, Stephen Morris wrote:
Sorry, this is my fault. I'm used to working with html mails where the response is black text and the quotes are colour coded for quotes from different responders, and I keep forgetting that environments that insist on using text only potentially don't have the colour coding to make things obvious.
I am seeing your HTML mail, and while it does highlight the edge of different sections, one huge slab of text makes it difficult to find new comments in the middle. And when replying to such things, quotes often get joined together that shouldn't.
It doesn't matter whether you post HTML or plain text, it helps things an awful lot when there's a gap between things. Just the same as it helps when people use paragraphs. I know someone on another list who can write three pages of text as one paragraph. My eyes run out of breath reading them.
My mail program does colour quotes different from new content, but with plain text mail all quotes are grey, new text is black. And if a poster didn't leave a gap between quotes and replies, often the reply is coloured the same as the quote. With HTML mail, it relies on the message styling itself. Yours are all black text, with coloured stripes running down the side.
That's interesting. In Thunderbird I'm seeing things differently, depending on who is reply to threads I'm seeing the entire quote with a black line down the side and text in black, for other people I see a coloured stripe and the text in blue. For the quotes you are replying to I see a dark blue stripe and light blue text with your response in black. What is also interesting is that for to portion of your response I'm replying to, in this composition that quote has a dark blue stripe and black text.
I think that's the difference between Thunderbird showing a plain text message vs. an HTML message. Since you send both, I can switch between which one is shown and see the difference as you describe. As someone else mentioned, with HTML messages Thunderbird can only show it as it was sent, so that's why you get black quoted text. With plain text, it identifies the quotes, colors the text differently, and creates the quote bars on the left with different colors.
Thankyou. I thought the colour coding was Thunderbirds html rendering and the version where everything was black was its text rendering.
regards, Steve
On Tue, 2024-12-03 at 09:15 +1100, Stephen Morris wrote:
That's interesting. In Thunderbird I'm seeing things differently, depending on who is reply to threads I'm seeing the entire quote with a black line down the side and text in black, for other people I see a coloured stripe and the text in blue. For the quotes you are replying to I see a dark blue stripe and light blue text with your response in black. What is also interesting is that for to portion of your response I'm replying to, in this composition that quote has a dark blue stripe and black text.
It's been a while since I've played with Thunderbird, and not with the latest version of it (obviously), but it was my understanding (after experimentation), that if:
Thunderbird receives some plain-text mail, it'll do a HTML rendition of it to display it, as it sees fit.
If it receives a HTML mail, the styling embedded in the message will dictate the formatting.
When it comes to replying to either of those, it's going to be a bit hit and miss. Some other mail client's idea of HTML mail may be different from Thunderbird's, and could be difficult for Thunderbird to determine quotes of quotes from other content.
If the HTML mail did the sensible thing and encased quoted text in blockquote, that identifies it as a quote. If the HTML mail just played silly-buggers with paragraphs, divs, etc, there's little clue as to whether that's quoted text or otherwise-styled content. It gets all the more messy if someone breaks into the middle of a slab of quoted text, to reply, but doesn't sufficiently create a clearly obvious break.
This chaotic mess is yet another reason, in the very long list of, why HTML is frowned on in certain circles (particularly mailing lists). For a long time there has not been a standard way to do HTML mail.
Plain text, on the other hand, has had some well established techniques. Quoted text gets a greater-then sign prefixing it (occasionally another symbol, but only the greater-than and pipe symbol gets wide recognition). Quotes of quotes get another one stacked against it, and so on, and so forth.
Any mail client that still gets that wrong, and many do, has bad coding. Evolution frequently does. I nearly always have to hand-edit quoted text to unscramble the mess it would otherwise leave.
Forté Inc's Agent (a Windows usenet client that also did mail), was the only one I ever came across that did it without any screw-ups. Ritlab's The Bat! was the only other mail program I used on Windows that could do email without looking like a dyslexic 10 year had coded the program.
The mail program, or you, should be able reflow the text beside the quote indicators without it stupidly scrambling the greater-than symbols into the middle of the text. Without mixing up the quotes of quotes, into a single generation quote, and into the replies... But so many of them are just utterly crap at doing their job.
Evolution appears to create all posts using HTML, but will convert it to plain text when sending it, if you set your options for plain text. You can see this in the manner that the editor works while typing in the message editor, and if you save a draft. Every now and then it'd do a spectacular cock-up where deleting one word would also remove everything else below it. Or you'd be unable to remove a line break that ended up in a stupid place.
Another pet hate with email over the last quarter century was mail clients that just can't get line wrapping right, and would turn messages into *this* kind of stupidly folded lines where there was a long line and a short overflow onto the next line, followed by another long line, ad infinitum, when quoting someone. It's a sodding bastard to read. Learn to code properly when you write email software, and either learn to properly wrap lines of text, or *don't* rewrap it.
On 3/12/24 13:09, Tim via users wrote:
On Tue, 2024-12-03 at 09:15 +1100, Stephen Morris wrote:
That's interesting. In Thunderbird I'm seeing things differently, depending on who is reply to threads I'm seeing the entire quote with a black line down the side and text in black, for other people I see a coloured stripe and the text in blue. For the quotes you are replying to I see a dark blue stripe and light blue text with your response in black. What is also interesting is that for to portion of your response I'm replying to, in this composition that quote has a dark blue stripe and black text.
It's been a while since I've played with Thunderbird, and not with the latest version of it (obviously), but it was my understanding (after experimentation), that if:
Thunderbird receives some plain-text mail, it'll do a HTML rendition of it to display it, as it sees fit.
If it receives a HTML mail, the styling embedded in the message will dictate the formatting.
When it comes to replying to either of those, it's going to be a bit hit and miss. Some other mail client's idea of HTML mail may be different from Thunderbird's, and could be difficult for Thunderbird to determine quotes of quotes from other content.
If the HTML mail did the sensible thing and encased quoted text in blockquote, that identifies it as a quote. If the HTML mail just played silly-buggers with paragraphs, divs, etc, there's little clue as to whether that's quoted text or otherwise-styled content. It gets all the more messy if someone breaks into the middle of a slab of quoted text, to reply, but doesn't sufficiently create a clearly obvious break.
This chaotic mess is yet another reason, in the very long list of, why HTML is frowned on in certain circles (particularly mailing lists). For a long time there has not been a standard way to do HTML mail.
Plain text, on the other hand, has had some well established techniques. Quoted text gets a greater-then sign prefixing it (occasionally another symbol, but only the greater-than and pipe symbol gets wide recognition). Quotes of quotes get another one stacked against it, and so on, and so forth.
Thankyou, I've seen those greater-than signs in a mail in this thread, but only in one and from what I see in mails on this list that I open I've only ever seen that once, since I installed F38 from scratch. I'm on F41 now and as I said that one mail is the first time I've seen that convention. I'm also on the Daily version of Thunderbird, so relative to what's in the Fedora repositories, I'm several versions ahead, so what I'm seeing is what other Thunderbird users will see when the repositories get to the version I'm on now, unless things have been proven to be defects.
Any mail client that still gets that wrong, and many do, has bad coding. Evolution frequently does. I nearly always have to hand-edit quoted text to unscramble the mess it would otherwise leave.
Forté Inc's Agent (a Windows usenet client that also did mail), was the only one I ever came across that did it without any screw-ups. Ritlab's The Bat! was the only other mail program I used on Windows that could do email without looking like a dyslexic 10 year had coded the program.
The mail program, or you, should be able reflow the text beside the quote indicators without it stupidly scrambling the greater-than symbols into the middle of the text. Without mixing up the quotes of quotes, into a single generation quote, and into the replies... But so many of them are just utterly crap at doing their job.
Evolution appears to create all posts using HTML, but will convert it to plain text when sending it, if you set your options for plain text. You can see this in the manner that the editor works while typing in the message editor, and if you save a draft. Every now and then it'd do a spectacular cock-up where deleting one word would also remove everything else below it. Or you'd be unable to remove a line break that ended up in a stupid place.
I'm not sure how Evolution does its thing as I've never used it. If I had the time to invest in setting up a mail system I would still be using Lotus Notes.
Another pet hate with email over the last quarter century was mail clients that just can't get line wrapping right, and would turn messages into *this* kind of stupidly folded lines where there was a long line and a short overflow onto the next line, followed by another long line, ad infinitum, when quoting someone. It's a sodding bastard to read. Learn to code properly when you write email software, and either learn to properly wrap lines of text, or *don't* rewrap it.
Yeah, I have seen a mail package do that, but I have also seen people deliberately format their mail that way and had to reformat the text when moving it out of the mail, eg: when moving the text into excel to server as instructions on what should be done in the form environment being exposed by the excel workbook.
regards, Steve
Stephen Morris:
I'm not sure how Evolution does its thing as I've never used it. If I had the time to invest in setting up a mail system I would still be using Lotus Notes.
Over the years I've used Amiga, Windows, Linux, Mac, and mainframes before the public internet (still have a few Data General punch cards, well the ones you filled in with lead pencil rather than punched holes with a tool). I've tried out a variety of email programs, often because the default application was absolutely crap at many of the things I've already brought up.
Out of what are probably the two current main ones on Linux, Evolution and Thunderbird, I found Evolution to be the least worst. I know, that's a terrible way to select a program to use. Thunderbird's slowness and what you see as you type isn't necessarily what you'll get when you're done, vetoes it for me. Likewise with it's modification of received mail to re-render it. I don't use KDE, I can't stand it, and I didn't like their email program when I tried it long ago.
Main criteria: * IMAP * Ease of navigating through folders and new mail * Not making reading a message difficult * Not absolutely crap at replying to messages * Not slow as mollasses bogging down the CPU * Not a complete pain to transition mail from an old install to a new system installation (mostly made easier by a local IMAP server and never using the mail client as local storage). * Able to turn off *helpful* features (continual background checking for new mail, mail filtering, unfettered downloading of things included in HTML mail, etc). * No hideously designed interfaces, which includes how you configure it, not just how you use it.
Tim:
Another pet hate with email over the last quarter century was mail clients that just can't get line wrapping right, and would turn messages into *this* kind of stupidly folded lines where there was a long line and a short overflow onto the next line
Yeah, I have seen a mail package do that, but I have also seen people deliberately format their mail that way and had to reformat the text when moving it out of the mail, eg: when moving the text into excel to server as instructions on what should be done in the form environment being exposed by the excel workbook.
It's due to incompetent programming. Email has two MAIN functions of reading and writing mail. Get THAT right before adding fancy features.
Much of that is caused by old habits of keeping messages under 80 columns wide, from the days of reading text using command line programs. And not knowing how to sensibly re-wrap text. Not helped by programs that decide to hard wrap text when they don't need to.
Few people read email on a console, most read it in a GUI, often with a proportional font (not the best choice for anything that typed code or figures). Email does NOT have a 80 column line length restriction. There are client and server programs that do have line length issues, but it tends be something more like 999 characters, or much greater (i.e. massively different than 80 characters).
So already 80 column pre-wrapped text can stay as it is, just becoming two or more characters longer by the greater-than quote prefixes, and still be perfectly readable by the vast majority of readers. Or could be sensibly re-flowed, but is often better not because the author may have hand-tabulated data, or copied and pasted some code.
Hand typed text certainly doesn't need to be stuffed up that way, either. I've seen clients that allow you type type longer than 80 character lines, then badly mangle it down to 80 characters afterwards. I seem to recall that was Microsoft's appalling Outlook Express (with a gazillion other stupidities built into it - let's execute attachments without due care, for instance, not even bothering to ask the recipient first).
On Wed, 2024-12-04 at 11:11 +1030, Tim via users wrote:
Out of what are probably the two current main ones on Linux, Evolution and Thunderbird, I found Evolution to be the least worst. I know, that's a terrible way to select a program to use. Thunderbird's slowness and what you see as you type isn't necessarily what you'll get when you're done, vetoes it for me. Likewise with it's modification of received mail to re-render it. I don't use KDE, I can't stand it, and I didn't like their email program when I tried it long ago.
I use Evolution on KDE and always have done. Can't get my head around Gnome, but that's another topic.
Main criteria: * IMAP * Ease of navigating through folders and new mail * Not making reading a message difficult * Not absolutely crap at replying to messages * Not slow as mollasses bogging down the CPU * Not a complete pain to transition mail from an old install to a new system installation (mostly made easier by a local IMAP server and never using the mail client as local storage). * Able to turn off *helpful* features (continual background checking for new mail, mail filtering, unfettered downloading of things included in HTML mail, etc). * No hideously designed interfaces, which includes how you configure it, not just how you use it.
One more: good support for MS Exchange, useful to those in corporate environments that require it. Maybe TBird has this as well, I wouldn't know, but it's worth mentioning.
Also, the principle developer Milan Crha is very responsive to questions and bug reports.
See https://lists.osuosl.org/mailman/listinfo/evolution-users
poc
On 4/12/24 21:03, Patrick O'Callaghan wrote:
On Wed, 2024-12-04 at 11:11 +1030, Tim via users wrote:
Out of what are probably the two current main ones on Linux, Evolution and Thunderbird, I found Evolution to be the least worst. I know, that's a terrible way to select a program to use. Thunderbird's slowness and what you see as you type isn't necessarily what you'll get when you're done, vetoes it for me. Likewise with it's modification of received mail to re-render it. I don't use KDE, I can't stand it, and I didn't like their email program when I tried it long ago.
I use Evolution on KDE and always have done. Can't get my head around Gnome, but that's another topic.
Main criteria: * IMAP * Ease of navigating through folders and new mail * Not making reading a message difficult * Not absolutely crap at replying to messages * Not slow as mollasses bogging down the CPU * Not a complete pain to transition mail from an old install to a new system installation (mostly made easier by a local IMAP server and never using the mail client as local storage). * Able to turn off *helpful* features (continual background checking for new mail, mail filtering, unfettered downloading of things included in HTML mail, etc). * No hideously designed interfaces, which includes how you configure it, not just how you use it.
One more: good support for MS Exchange, useful to those in corporate environments that require it. Maybe TBird has this as well, I wouldn't know, but it's worth mentioning.
I have to use Exchange/Outlook at work and I hate it. My view has always been that the Exchange/Outlook combination was Microsoft's attempt to replicate Lotus Notes and it has always been a dismal failure. Also, in my view, the organisation made a huge mistake in moving off Lotus Notes.
regards, Steve
Also, the principle developer Milan Crha is very responsive to questions and bug reports.
Seehttps://lists.osuosl.org/mailman/listinfo/evolution-users
poc
On Wed, 04 Dec 2024 11:11:33 +1030 "Tim via users" users@lists.fedoraproject.org wrote:
Stephen Morris:
I'm not sure how Evolution does its thing as I've never used it. If I had the time to invest in setting up a mail system I would still be using Lotus Notes.
Over the years I've used Amiga, Windows, Linux, Mac, and mainframes before the public internet (still have a few Data General punch cards, well the ones you filled in with lead pencil rather than punched holes with a tool). I've tried out a variety of email programs, often because the default application was absolutely crap at many of the things I've already brought up.
Out of what are probably the two current main ones on Linux, Evolution and Thunderbird, I found Evolution to be the least worst. I know, that's a terrible way to select a program to use. Thunderbird's slowness and what you see as you type isn't necessarily what you'll get when you're done, vetoes it for me. Likewise with it's modification of received mail to re-render it. I don't use KDE, I can't stand it, and I didn't like their email program when I tried it long ago.
Main criteria:
- IMAP
- Ease of navigating through folders and new mail
- Not making reading a message difficult
- Not absolutely crap at replying to messages
- Not slow as mollasses bogging down the CPU
- Not a complete pain to transition mail from an old install to a new system installation (mostly made easier by a local IMAP server and never using the mail client as local storage).
- Able to turn off *helpful* features (continual background checking for new mail, mail filtering, unfettered downloading of things included in HTML mail, etc).
- No hideously designed interfaces, which includes how you configure it, not just how you use it.
Doesn't Claws Mail satisfy all this criteria? To me, undoubtedly.
On Wed, 2024-12-04 at 11:04 +0000, Bob Marčan via users wrote:
Doesn't Claws Mail satisfy all this criteria? To me, undoubtedly.
I did look at it, ages ago. And I do mean a very long time ago (this will be in the pre-Fedora era of Red Hat Linux). I didn't like it at the time, for reasons I've mostly forgotten. It's probably different by now. It's been a while since I went mail app comparing, you can go down a rabbit hole of spending a lot of time trying things out trying to weigh up which 6.5 out of 10 stars mail program you'd prefer to use, instead of just actually doing whatever work you were intending to do.
Buying a new PC was similarly hair-pulling exercises. Eventually it becomes a what seems best in my price range decision. I was quite pleased with my last one (sufficiently processing ability powerful, quiet, quite low power consumption).
One thing that springs back to mind, looking at a screen shot, was cluttered display. I never liked things with clicky-things here, there, and everywhere. All in the top of the window is where I want them. Apart from the hunting around and visual distractions, it uses up valuable screen screen space. I detest the "looking through a keyhole" view that some interfaces create.
Back then I was using a CRT monitor with probably 800 vertical pixels, so wasting screen space was a really big problem. Now I have a bit more resolution, but not masses more. It's 1920 by 1080. And emailing is often a side-by-side thing with a web browser alongside it.
Interface design, and how painful it was to write and reply, were my chief go/no-go decision making back then.
Maildir became another criteria (with any mail client). Massively huge single mbox files become a serious speed and stability problem. That's been my actual experience, rather than just a theoretical thing.
Even though I use dovecot as a local IMAP mail server, mail clients like to locally cache things. Every time you fetched new mail, mbox would be a laborious update before you could do anything. Evolution was awful at that, Thunderbird was even worse. Switching to maildir made a huge improvement. I had to forcekill Thunderbird many times.
On 4/12/24 11:41, Tim via users wrote:
Stephen Morris:
I'm not sure how Evolution does its thing as I've never used it. If I had the time to invest in setting up a mail system I would still be using Lotus Notes.
Over the years I've used Amiga, Windows, Linux, Mac, and mainframes before the public internet (still have a few Data General punch cards, well the ones you filled in with lead pencil rather than punched holes with a tool). I've tried out a variety of email programs, often because the default application was absolutely crap at many of the things I've already brought up.
Over the years I've used all of those systems as well except for Mac's, and some of the email packages on those systems were horrendous.
Out of what are probably the two current main ones on Linux, Evolution and Thunderbird, I found Evolution to be the least worst. I know, that's a terrible way to select a program to use. Thunderbird's slowness and what you see as you type isn't necessarily what you'll get when you're done, vetoes it for me. Likewise with it's modification of received mail to re-render it. I don't use KDE, I can't stand it, and I didn't like their email program when I tried it long ago.
Of all the email packages I've tried I thought Thunderbird was the easiest to use and set up. I'm not impressed with how many threads it runs in the background, but as those threads are dormant they are not really consuming any significant amount of resources.
Main criteria:
- IMAP
- Ease of navigating through folders and new mail
- Not making reading a message difficult
- Not absolutely crap at replying to messages
- Not slow as mollasses bogging down the CPU
- Not a complete pain to transition mail from an old install to a new system installation (mostly made easier by a local IMAP server and never using the mail client as local storage).
- Able to turn off *helpful* features (continual background checking for new mail, mail filtering, unfettered downloading of things included in HTML mail, etc).
- No hideously designed interfaces, which includes how you configure it, not just how you use it.
Thunderbird also provides most of those features, and I have Thunderbird set up to display the most recent mail, be it in the mailbox or folders it moved to by filters, at the top of the list, and with unread mail being shown in bold text they stand out.
regards, Steve
On 1/12/24 17:04, Tim via users wrote:
The issue with this is what you have shown is what Thunderbird used to always show to identify who said what which I find extremely useful to understand response history and the context of replies.
Again, look through your preferences, and see if there are options about hiding/muting/suppressing quoted text. Various mail programs have that option.
I have Thunderbird configured to automatically quote the original message being replied to.
regards, Steve
Tim:
Again, look through your preferences, and see if there are options about hiding/muting/suppressing quoted text. Various mail programs have that option.
Stephen Morris:
I have Thunderbird configured to automatically quote the original message being replied to.
Since you mentioned not seeing some aspects of people's messages *until* you replied to them, I was guessing that your mail program was suppressing them.
On 2/12/24 22:22, Tim via users wrote:
Tim:
Again, look through your preferences, and see if there are options about hiding/muting/suppressing quoted text. Various mail programs have that option.
Stephen Morris:
I have Thunderbird configured to automatically quote the original message being replied to.
Since you mentioned not seeing some aspects of people's messages *until* you replied to them, I was guessing that your mail program was suppressing them.
I didn't remember ever seeing an option in Thunderbird to suppress those messages but that was a reasonable check to do.
regards, Steve
On Tue, 2024-12-03 at 09:19 +1100, Stephen Morris wrote:
I didn't remember ever seeing an option in Thunderbird to suppress those messages but that was a reasonable check to do.
Apparently, at some stage Thunderbird had an option for that (it called it QuoteCollapse), and I saw a comment that it may have come back. Dunno if they meant internally, or as an add-on. You could search your settings for that, or in the about:config under-the-bonnet tweaking options.
Other names for the same feature: suppress, mask, mute, hide
Some mail clients have options for hiding quotes, and another separately, for signatures (the boilerplate text, not PGP signatures). I only see hiding them causing problems, hiding bits of messages that maybe they shouldn't.
I remember one client that had a dimming option, it dulled out the quoted text (more than the mid-grey Evolution is doing for me, now, but I can adjust that), so that new text was *much* more prominent than quoted text, but all of it was still readable. And you could flick it on and off without reconfiguring the client. A useful feature, so long as it could actually tell the difference between old quotes and new content.
Other's simply checkboarded odd and even generations of quotes between two different colour schemes. Too-many colours all over the place can be a bit overwhelming, but limiting it to three wasn't too bad (new text, one set of quotes, another set of quotes). Though people should really edit out very old message content.
On 3/12/24 13:30, Tim via users wrote:
I didn't remember ever seeing an option in Thunderbird to suppress those messages but that was a reasonable check to do.
Apparently, at some stage Thunderbird had an option for that (it called it QuoteCollapse), and I saw a comment that it may have come back. Dunno if they meant internally, or as an add-on. You could search your settings for that, or in theabout:config under-the-bonnet tweaking options.
I'm not sure how to enter an about:config as there is no command line to enter that into to find out if there is an internal option.
regards, Steve
On Wed, 2024-12-04 at 08:31 +1100, Stephen Morris wrote:
I'm not sure how to enter an about:config as there is no command line to enter that into to find out if there is an internal option.
No, I couldn't figure out the trick, either. It used to be doable without too much tearing out your hair. But the various internet searches I did only had hints for different versions of Thunderbird than I had to play with. None of their /look in this menu/ instructions were right.
On 12/3/24 1:31 PM, Stephen Morris wrote:
On 3/12/24 13:30, Tim via users wrote:
I didn't remember ever seeing an option in Thunderbird to suppress those messages but that was a reasonable check to do.
Apparently, at some stage Thunderbird had an option for that (it called it QuoteCollapse), and I saw a comment that it may have come back. Dunno if they meant internally, or as an add-on. You could search your settings for that, or in theabout:config under-the-bonnet tweaking options.
I'm not sure how to enter an about:config as there is no command line to enter that into to find out if there is an internal option.
A quick Google search leads directly to the mozilla help documentation: https://support.mozilla.org/en-US/kb/config-editor
At the bottom of the general settings, there's a button for the config editor. I've used it before, but forgot that's where it was.
On 4/12/24 21:23, Samuel Sieb wrote:
On 12/3/24 1:31 PM, Stephen Morris wrote:
On 3/12/24 13:30, Tim via users wrote:
I didn't remember ever seeing an option in Thunderbird to suppress those messages but that was a reasonable check to do.
Apparently, at some stage Thunderbird had an option for that (it called it QuoteCollapse), and I saw a comment that it may have come back. Dunno if they meant internally, or as an add-on. You could search your settings for that, or in theabout:config under-the-bonnet tweaking options.
I'm not sure how to enter an about:config as there is no command line to enter that into to find out if there is an internal option.
A quick Google search leads directly to the mozilla help documentation: https://support.mozilla.org/en-US/kb/config-editor
At the bottom of the general settings, there's a button for the config editor. I've used it before, but forgot that's where it was.
Thankyou. I have forgotten it was there and I'd missed it when trolling through the settings.
regards, Steve
On 5/12/24 09:28, Stephen Morris wrote:
On 4/12/24 21:23, Samuel Sieb wrote:
On 12/3/24 1:31 PM, Stephen Morris wrote:
On 3/12/24 13:30, Tim via users wrote:
I didn't remember ever seeing an option in Thunderbird to suppress those messages but that was a reasonable check to do.
Apparently, at some stage Thunderbird had an option for that (it called it QuoteCollapse), and I saw a comment that it may have come back. Dunno if they meant internally, or as an add-on. You could search your settings for that, or in theabout:config under-the-bonnet tweaking options.
I'm not sure how to enter an about:config as there is no command line to enter that into to find out if there is an internal option.
A quick Google search leads directly to the mozilla help documentation: https://support.mozilla.org/en-US/kb/config-editor
At the bottom of the general settings, there's a button for the config editor. I've used it before, but forgot that's where it was.
Thankyou. I have forgotten it was there and I'd missed it when trolling through the settings.
Not unexpectedly there wasn't an about:config option, so I checked for config and in the results there doesn't appear to be an option to suppress quoting, so I guess the option to automatically quote the original text in replies is active.
regards, Steve
regards, Steve
On 11/27/24 1:57 PM, Stephen Morris wrote:
On 27/11/24 12:48, Tim via users wrote:
On Wed, 2024-11-27 at 08:46 +1100, Stephen Morris wrote:
This is true, and I know this can be done under DNF, but I'm using that to decide whether it is worth doing the upgrade in terms of the volume of updates that will be put on. If there are only 3 updates to put on then its not necessarily worth putting on the updates yet.
So, just do "sudo dnf update" and when it comes back with the results, look at them, THEN hit Y or N for yes or no to install them.
People who preload the install/update/erase command line with yes are just asking for trouble. If something goes rogue, a command may install or remove hundreds of packages, leaving you with a borked system.
That's another bit of bad advice various websites offer.
I never use -y. I know I could issue dnf upgrade and do the same thing, I've just been playing around with dnf check-upgrade. I could be wrong but dnf check-upgrade seems to me to be new with dnf5, so presumably it has been added for a reason, so is there checking done that is not done with dnf upgrade?
It's most likely intended to be used by shell scripts.
On 28/11/24 20:11, Samuel Sieb wrote:
On 11/27/24 1:57 PM, Stephen Morris wrote:
On 27/11/24 12:48, Tim via users wrote:
On Wed, 2024-11-27 at 08:46 +1100, Stephen Morris wrote:
This is true, and I know this can be done under DNF, but I'm using that to decide whether it is worth doing the upgrade in terms of the volume of updates that will be put on. If there are only 3 updates to put on then its not necessarily worth putting on the updates yet.
So, just do "sudo dnf update" and when it comes back with the results, look at them, THEN hit Y or N for yes or no to install them.
People who preload the install/update/erase command line with yes are just asking for trouble. If something goes rogue, a command may install or remove hundreds of packages, leaving you with a borked system.
That's another bit of bad advice various websites offer.
I never use -y. I know I could issue dnf upgrade and do the same thing, I've just been playing around with dnf check-upgrade. I could be wrong but dnf check-upgrade seems to me to be new with dnf5, so presumably it has been added for a reason, so is there checking done that is not done with dnf upgrade?
It's most likely intended to be used by shell scripts.
It could be, thankyou.
regards, Steve
On Tue, 2024-11-26 at 09:02 +1100, Stephen Morris wrote:
I assumed that when dnf retrieved it's metadata it was updating the standard system locations which a normal user can't update, I didn't know it was caching the data elsewhere.
If you haven't given the magic password, no command is going to have the authority to change non-personal-user data.
It's for these kinds of situations that there is a background task that pops alerts up through the taskbar to let people know there's updates waiting (if you want to use it). It does have some elevated privileges.
On 26/11/24 12:10, Tim via users wrote:
On Tue, 2024-11-26 at 09:02 +1100, Stephen Morris wrote:
I assumed that when dnf retrieved it's metadata it was updating the standard system locations which a normal user can't update, I didn't know it was caching the data elsewhere.
If you haven't given the magic password, no command is going to have the authority to change non-personal-user data.
It's for these kinds of situations that there is a background task that pops alerts up through the taskbar to let people know there's updates waiting (if you want to use it). It does have some elevated privileges.
Yes, and I have never liked its functionality but put up with it because I couldn't figure out how to disable it, although until now I didn't know that what I don't like about it is being brought about by the nature of the update repositories potentially. What I didn't like with it was it pops up a message about updates being available and when I look at what it is saying, it is notifying about updates for the packages I have already just put on via DNF.
regards, Steve
On Wed, 2024-11-27 at 08:51 +1100, Stephen Morris wrote:
On 26/11/24 12:10, Tim via users wrote:
On Tue, 2024-11-26 at 09:02 +1100, Stephen Morris wrote:
I assumed that when dnf retrieved it's metadata it was updating the standard system locations which a normal user can't update, I didn't know it was caching the data elsewhere.
If you haven't given the magic password, no command is going to have the authority to change non-personal-user data.
It's for these kinds of situations that there is a background task that pops alerts up through the taskbar to let people know there's updates waiting (if you want to use it). It does have some elevated privileges.
Yes, and I have never liked its functionality but put up with it because I couldn't figure out how to disable it, although until now I didn't know that what I don't like about it is being brought about by the nature of the update repositories potentially. What I didn't like with it was it pops up a message about updates being available and when I look at what it is saying, it is notifying about updates for the packages I have already just put on via DNF.
You can just remove it. I only ever use dnf so the GUI updater just gets in the way.
poc
On 27/11/24 09:17, Patrick O'Callaghan wrote:
On Wed, 2024-11-27 at 08:51 +1100, Stephen Morris wrote:
On 26/11/24 12:10, Tim via users wrote:
On Tue, 2024-11-26 at 09:02 +1100, Stephen Morris wrote:
I assumed that when dnf retrieved it's metadata it was updating the standard system locations which a normal user can't update, I didn't know it was caching the data elsewhere.
If you haven't given the magic password, no command is going to have the authority to change non-personal-user data.
It's for these kinds of situations that there is a background task that pops alerts up through the taskbar to let people know there's updates waiting (if you want to use it). It does have some elevated privileges.
Yes, and I have never liked its functionality but put up with it because I couldn't figure out how to disable it, although until now I didn't know that what I don't like about it is being brought about by the nature of the update repositories potentially. What I didn't like with it was it pops up a message about updates being available and when I look at what it is saying, it is notifying about updates for the packages I have already just put on via DNF.
You can just remove it. I only ever use dnf so the GUI updater just gets in the way.
It may already be removed, I'm not sure, but I haven't noticed its popup since upgrading to F41.
regards, Steve
poc
-
Bob Marčan -
Go Canes -
Jonathan Billings -
Patrick O'Callaghan -
Samuel Sieb -
Stephen Morris -
Tim -
Will McDonald