There was a discussion earlier today on the fedora-list related with the fact that cdrecord doesn't have a man page. The person who made the question didn't know that cdrecord is just a symlink to wodim, so that man wodim would have given the man page he was looking for.
The thing is that the discussion started flipping towards why, if there is a binary symlink, there isn't a man page symlink? This could, in the case of cdrecord and wodim, or mkisofs and genisofs, end in bugtrack tickets for those packages, but I am more interested in a general solution, for all packages that have binary symlinks.
So, shouldn't there be something in respect of this issue in the packaging policy?
Martin Marques wrote:
This could, in the case of cdrecord and wodim, or mkisofs and genisofs, end in bugtrack tickets for those packages, but I am more interested in a general solution, for all packages that have binary symlinks.
Skipping the larger policy issue, I did file a bug (with a patch) about adding the man page symlinks for cdrkit.
https://bugzilla.redhat.com/show_bug.cgi?id=296611
On Wednesday 19 September 2007, Martin Marques wrote:
The thing is that the discussion started flipping towards why, if there is a binary symlink, there isn't a man page symlink? This could, in the case of cdrecord and wodim, or mkisofs and genisofs, end in bugtrack tickets for those packages, but I am more interested in a general solution, for all packages that have binary symlinks.
So, shouldn't there be something in respect of this issue in the packaging policy?
I support that (or the more or less AFAIK equivalent 'echo ".so man1/wodim.1"
cdrecord.1' approach as an alternative to symlinks). Could you write a
draft for that? I can take it onwards from there in the packaging committee.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Wed, 19 Sep 2007 21:44:16 +0300 Ville Skyttä ville.skytta@iki.fi wrote:
On Wednesday 19 September 2007, Martin Marques wrote:
The thing is that the discussion started flipping towards why, if there is a binary symlink, there isn't a man page symlink? This could, in the case of cdrecord and wodim, or mkisofs and genisofs, end in bugtrack tickets for those packages, but I am more interested in a general solution, for all packages that have binary symlinks.
So, shouldn't there be something in respect of this issue in the packaging policy?
I support that (or the more or less AFAIK equivalent 'echo ".so man1/wodim.1"
cdrecord.1' approach as an alternative to symlinks). Could you write a
draft for that? I can take it onwards from there in the packaging committee.
I would suggest creating a cdrecord(1) man page that says "This is no more. See wodim(1), which replaces it." Put that cdrecord(1) man page into the wodim package. This way, we teach people that there's been a change and we don't get "Um, when I type 'man cdrecord' I get a man page for something totally different; what gives?" kinds of questions/bug reports/requests.
Of course, adjust the verbage appropriately in the "this command went bye-bye" man pages. - -- Lamont Peterson lamont@gurulabs.com Senior Instructor Guru Labs, L.C. [ http://www.GuruLabs.com/ ]
NOTE: All messages from this email address should be digitally signed with my 0xDC0DD409 GPG key. It is available on the pgp.mit.edu keyserver as well as other keyservers that sync with MIT's.
Lamont Peterson escribió:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Wed, 19 Sep 2007 21:44:16 +0300 Ville Skyttä ville.skytta@iki.fi wrote:
On Wednesday 19 September 2007, Martin Marques wrote:
The thing is that the discussion started flipping towards why, if there is a binary symlink, there isn't a man page symlink? This could, in the case of cdrecord and wodim, or mkisofs and genisofs, end in bugtrack tickets for those packages, but I am more interested in a general solution, for all packages that have binary symlinks.
So, shouldn't there be something in respect of this issue in the packaging policy?
I support that (or the more or less AFAIK equivalent 'echo ".so man1/wodim.1"
cdrecord.1' approach as an alternative to symlinks). Could you write a
draft for that? I can take it onwards from there in the packaging committee.
I would suggest creating a cdrecord(1) man page that says "This is no more. See wodim(1), which replaces it." Put that cdrecord(1) man page into the wodim package. This way, we teach people that there's been a change and we don't get "Um, when I type 'man cdrecord' I get a man page for something totally different; what gives?" kinds of questions/bug reports/requests.
If they read the man they would note that:
NOTE There may be similarities and differences between this program and other disk recording application(s). See the CREDITS and AUTHORS sec- tions below to learn about the origin of wodim.
And later that:
This application is a spinoff from the original program "cdrecord" delivered in the cdrtools package [1] created by Joerg Schilling, who deserves the most credits for its success. However, he is not involved into the development of this spinoff and therefore he shall not be made responsible for any problem caused by it. Do not refer to this applica- tion as "cdrecord", do not try to get support for wodim by contacting the original authors.
So, better would be to teach users to read the manual.
This is obviously IMHO.
Lamont Peterson lamont@gurulabs.com Senior Instructor Guru Labs, L.C. [ http://www.GuruLabs.com/ ]
NOTE: All messages from this email address should be digitally signed with my 0xDC0DD409 GPG key. It is available on the pgp.mit.edu keyserver as well as other keyservers that sync with MIT's. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (GNU/Linux)
iD8DBQFG8Xon+YBsl9wN1AkRAhdtAJ43J8X8zY++WFz3J/yP2rp5zNhcUwCgpbtn YHhhBo5yl29FWGpPCR8RDTc= =KqqQ -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Wed, 19 Sep 2007 17:34:27 -0300 Martin Marques martin@bugs.unl.edu.ar wrote:
Lamont Peterson escribió:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Wed, 19 Sep 2007 21:44:16 +0300 Ville Skyttä ville.skytta@iki.fi wrote:
On Wednesday 19 September 2007, Martin Marques wrote:
The thing is that the discussion started flipping towards why, if there is a binary symlink, there isn't a man page symlink? This could, in the case of cdrecord and wodim, or mkisofs and genisofs, end in bugtrack tickets for those packages, but I am more interested in a general solution, for all packages that have binary symlinks.
So, shouldn't there be something in respect of this issue in the packaging policy?
I support that (or the more or less AFAIK equivalent 'echo ".so man1/wodim.1"
cdrecord.1' approach as an alternative to symlinks). Could you write a
draft for that? I can take it onwards from there in the packaging committee.
I would suggest creating a cdrecord(1) man page that says "This is no more. See wodim(1), which replaces it." Put that cdrecord(1) man page into the wodim package. This way, we teach people that there's been a change and we don't get "Um, when I type 'man cdrecord' I get a man page for something totally different; what gives?" kinds of questions/bug reports/requests.
If they read the man they would note that:
NOTE There may be similarities and differences between this program and other disk recording application(s). See the CREDITS and AUTHORS sec- tions below to learn about the origin of wodim.
And later that:
This application is a spinoff from the original program "cdrecord" delivered in the cdrtools package [1] created by Joerg Schilling, who deserves the most credits for its success. However, he is not involved into the development of this spinoff and therefore he shall not be made responsible for any problem caused by it. Do not refer to this applica- tion as "cdrecord", do not try to get support for wodim by contacting the original authors.
So, better would be to teach users to read the manual.
This is obviously IMHO.
I honestly do not understand how what you are saying here helps in any way to answer the question being discussed in this thread. Of course it's best for the users to read the manual. That's not the point. We're talking about how to help people find the manual when the name has changed from something (like cdrecord, which prompted this thread in the first place) that has been around a long time.
Of course I saw the item you quoted/mentioned in the wodim(1) man page. It doesn't help anyone who doesn't know that wodim is what they want because cdrecord no longer exists on their systems (in the distribution).
If I've missed something here, help me understand it. - -- Lamont Peterson lamont@gurulabs.com Senior Instructor Guru Labs, L.C. [ http://www.GuruLabs.com/ ]
NOTE: All messages from this email address should be digitally signed with my 0xDC0DD409 GPG key. It is available on the pgp.mit.edu keyserver as well as other keyservers that sync with MIT's.
Lamont Peterson escribió:
I honestly do not understand how what you are saying here helps in any way to answer the question being discussed in this thread. Of course it's best for the users to read the manual. That's not the point. We're talking about how to help people find the manual when the name has changed from something (like cdrecord, which prompted this thread in the first place) that has been around a long time.
Of course I saw the item you quoted/mentioned in the wodim(1) man page. It doesn't help anyone who doesn't know that wodim is what they want because cdrecord no longer exists on their systems (in the distribution).
OK, better: add at the beginning of the wodim man page that "wodim is a fork of cdrecord". How about that?
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Wed, 19 Sep 2007 18:38:41 -0300 Martin Marques martin@bugs.unl.edu.ar wrote:
Lamont Peterson escribió:
I honestly do not understand how what you are saying here helps in any way to answer the question being discussed in this thread. Of course it's best for the users to read the manual. That's not the point. We're talking about how to help people find the manual when the name has changed from something (like cdrecord, which prompted this thread in the first place) that has been around a long time.
Of course I saw the item you quoted/mentioned in the wodim(1) man page. It doesn't help anyone who doesn't know that wodim is what they want because cdrecord no longer exists on their systems (in the distribution).
OK, better: add at the beginning of the wodim man page that "wodim is a fork of cdrecord". How about that?
I think the wodim(1) man page already does an adequate job of explaining this, though I do like your idea of a very concise piece of text at the beginning of the man page, just to make things more clear.
Still, the issue we're discussing in this thread isn't about problems with the wodim(1) man page. It's about the problem that there is no cdrecord(1) man page anymore. Since there is no cdrecord(1) man page, running "man cdrecord" just like users may have been doing for 5-10 years now, will not work.
However, they still have a "cdrecord" command (as a symlink to "wodim"), so they would now be frustrated and not know where to look for the "missing man page to cdrecord".
So, what this thread was trying to discuss was the idea that we need to do something for commands that still exist as symlinks as a convenience for users who are used to those commands and for scripts that might be using them, too.
My suggestion was to not create a symlink to, or a copy of, the wodim(1) man page in order to make "man cdrecord" work again, but to instead create a bona-fide cdrecord(1) man page with content such as "This isn't the command you want, anymore. You now want wodim(1)". - -- Lamont Peterson lamont@gurulabs.com Senior Instructor Guru Labs, L.C. [ http://www.GuruLabs.com/ ]
NOTE: All messages from this email address should be digitally signed with my 0xDC0DD409 GPG key. It is available on the pgp.mit.edu keyserver as well as other keyservers that sync with MIT's.
Lamont Peterson escribió:
My suggestion was to not create a symlink to, or a copy of, the wodim(1) man page in order to make "man cdrecord" work again, but to instead create a bona-fide cdrecord(1) man page with content such as "This isn't the command you want, anymore. You now want wodim(1)".
Let's see if I can get my ideas out. :-)
There is no cdrecord, and won't be any in future versions of Fedora. What there is is a symlink to wodim, just because wodim is a fork of cdrecord.
If I do man cdrecord and end up seeing the wodim man page, I would first nod my head, and then say "what's this wodim?" (maybe even ask a fellow linux friend) and then read the man.
To end this: why create a man page for something that doesn't exist, and maybe in a future version of fedora not even the symlink will be left (maybe yes, maybe not).
My opinion, link the manuals, and the info pages if they exist.
Lamont Peterson wrote:
My suggestion was to not create a symlink to, or a copy of, the wodim(1) man page in order to make "man cdrecord" work again, but to instead create a bona-fide cdrecord(1) man page with content such as "This isn't the command you want, anymore. You now want wodim(1)".
Unless there is a good reason to remove the cdrecord (and other old command name) symlinks, I would hope that they stick around. The name wodim (and the other new command names) is not nearly as useful as the more generic cdrecord is.
It's also my opinion that the man pages should remain available under the old (sane) names as well. Whether that's via a symlink or ".so man1/wodim.1" isn't terribly important to me. If there's a compelling reason to prefer one over the other, then let that be suggested in whatever packaging guidelines may arise from this discussion.