On Tue, Apr 3, 2012 at 3:40 PM, Chris Leonard cjlhomeaddress@gmail.com wrote:
On Tue, Apr 3, 2012 at 6:25 AM, Mikus Grinbergs mikus@bga.com wrote:
Peter wrote
We're glad of the testing but ultimately if you need a perfectly stable platform I suggest you stick with 11.3 for the moment.
I'm a volunteer. I have no __need__ to post here (nor to use the XO).
Which would you prefer - to have me keep quiet about what I encounter - or may I share my experiences without being lectured about __need__ ?
Mikus, You have a well earned reputation as a trailblazer in XO testing and it is a very valuable role that you play in pushing the envelope forward with the very latest builds often within minutes of their release. Please continue this excellent work.
I don't disagree but ultimately if there's not appropriate bug reports with all the information needed we can't fix the problems. So ultimately I appreciate his bug reports but with a non standard configuration we need detailed reports to have that information be useful for those people that actually have to deal with the information and fix the problem.
Peter
I don't disagree but ultimately if there's not appropriate bug reports with all the information needed we can't fix the problems. So ultimately I appreciate his bug reports but with a non standard configuration we need detailed reports to have that information be useful for those people that actually have to deal with the information and fix the problem.
I've already stopped writing bug reports against Fedora - I've concluded their requirement that the submitter describe "how to reproduce the bug" is to me not worth my time and effort. Fedora is *their* product - what does it matter to me whether it retains bugs ?
You would like "all the information needed". I thought I posted to the effect that when an XO-1 with 12.1.0 build os6 is booted, nothing gets defined in the /dev tree against which a 'mount' command could be issued in order to access the content of a partition on the SD card plugged into the XO-1 external slot (beneath the screen). It is not clear to me what the "we need detailed reports" are that would allow you to discover whether such a problem exists on the XO-1 you use.
Note that for my purposes I was able to bypass this problem. What I was not able to ignore was the os6 (on XO-1) reaction to having an entry in /etc/fstab referencing this undefined (on os6) storage device. But I agree that having such an entry in /etc/fstab is "non-standard" -- from now on, I will make sure to NEVER for 12.1.0 to mention any problems I have with whatever I do that is "non-standard".
mikus
On Tue, Apr 3, 2012 at 8:02 PM, Mikus Grinbergs mikus@bga.com wrote:
I don't disagree but ultimately if there's not appropriate bug reports with all the information needed we can't fix the problems. So ultimately I appreciate his bug reports but with a non standard configuration we need detailed reports to have that information be useful for those people that actually have to deal with the information and fix the problem.
I've already stopped writing bug reports against Fedora - I've concluded their requirement that the submitter describe "how to reproduce the bug" is to me not worth my time and effort. Fedora is *their* product - what does it matter to me whether it retains bugs ?
You would like "all the information needed". I thought I posted to the effect that when an XO-1 with 12.1.0 build os6 is booted, nothing gets defined in the /dev tree against which a 'mount' command could be issued in order to access the content of a partition on the SD card plugged into the XO-1 external slot (beneath the screen). It is not clear to me what the "we need detailed reports" are that would allow you to discover whether such a problem exists on the XO-1 you use.
Note that for my purposes I was able to bypass this problem. What I was not able to ignore was the os6 (on XO-1) reaction to having an entry in /etc/fstab referencing this undefined (on os6) storage device. But I agree that having such an entry in /etc/fstab is "non-standard" -- from now on, I will make sure to NEVER for 12.1.0 to mention any problems I have with whatever I do that is "non-standard".
Mikus, you misinterpret what I say, I have no problems with non-standard issues and reports, but you need to understand to be able fix the problem I need to know how I can recreate the problem or what the specifics of the problem are it makes it hard for me to work out where it needs to be fixed. I've already looked at the SD card issue and recreated the problem for example. You mentioned you'd already created a ticket for this, do you have the number of the ticket?
In the case of the /etc/fstab if it's a file system required to boot it will bork, if it's not it should but then with all the changes between f14 -> f17 for core boot process I'm not surprised if there's been a change that hasn't been documented. I'll try to recreate it as well. Tickets help me track the problems so it doesn't get lost.
Peter
In the case of the /etc/fstab if it's a file system required to boot it will bork
I don't mind it borking. But the system offered me the choice of "continue". When I took that choice, the system returned to the SAME bork. I would have preferred it if "continue" had meant "skip this".
You mentioned you'd already created a ticket for this, do you have the number of the ticket?
I wrote ticket # 11719 for a different problem - which appears to exist in all XO-1 12.1.0 builds (I believe I saw it before os3; it definitely exists in os5 and os6). [Whereas the "no device for the XO-1 SD card (beneath the display) problem only showed up with XO-1 build os6.]
Ticket # 11719 says that for XO-1 12.1.0 builds, nothing exists in the /dev tree which can be used in a 'mount' command to access the files which OFW uses to boot the XO-1. [These files exist in a separate partition on NAND; when things work correctly that partition gets mounted onto /bootpart -- then those files can be accessed/manipulated (e.g., via manual commands) from the running XO-1 system.]
My own interest is in customizing the olpc.fth file that OFW uses when it boots the XO-1. But the same problem arises if the user performs 'yum upgrade kernel' (assuming that a newer kernel version has been placed in the olpc-f17-xo1 repo). With the # 11719 problem present in XO-1 12.1.0, the output of the 'yum upgrade' (which got placed into /boot) CAN'T be copied onto /bootpart (because there is no filesystem mounted on /bootpart). The result is that the XO-1 12.1.0 system ignores attempts to upgrade its boot process, and continues to use those "boot files" which were created when the XO-1 12.1.0 build was originally compiled.
mikus
p.s. Please let me repeat:
It is not clear to me what the "we need detailed reports" are that would allow you to discover whether such a problem exists on the XO-1 you use.
-
Mikus Grinbergs -
Peter Robinson