Re: On the subject of watchfiles (was Re: Proposed MBF: Debian upstream version higher than watch file-reported upstream version)
Andreas Tille dijo [Tue, Feb 26, 2008 at 08:21:47PM +0100]: > >Heh, start a bit earlier (think Ruby)... Educate maintainers to > >release proper .tar.gz, not braindead .gem packages containing the > >equivalent to an orig.tar.gz (but created due to a nice > >don't-ask-me-why-that's-not-properly-implemented bug in December 31, > >1969)... And then complaining if you are distributing in stable > >anything older than their nightly checkouts. > > > >Yes, Perl and the CPAN rock my world, although my programming is > >nowadays mostly Ruby-based. The Ruby general mindset is WAY inferior. > > What? > I admit I would have been able to parse the contents of your mail > with the same success if you would have written in Spanish. :) > Prehaps it is me who had to get up 4:20 this morning (so I started > *really* early ;-) ) - but I do not even understand whether this is pro or > contra proper watch files. Sorry, I'm a bit incoherent as well - due to all kind of unrelated events :) I'm completely pro-watch. It is fundamental to the way pkg-perl works. As said IIRC by Tincho, there is no way to keep track of over 670 packages without automated tools. What really brings me down is that this is impossible in communities such as the Ruby one, which uses a incoherent and brain-dead packaging format, and insists on shoving it on distributions' throats. Bah. -- Gunnar Wolf - [EMAIL PROTECTED] - (+52-55)5623-0154 / 1451-2244 PGP key 1024D/8BB527AF 2001-10-23 Fingerprint: 0C79 D2D1 2C4E 9CE4 5973 F800 D80E F35A 8BB5 27AF -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: On the subject of watchfiles (was Re: Proposed MBF: Debian upstream version higher than watch file-reported upstream version)
On Tue, 26 Feb 2008, Gunnar Wolf wrote: Heh, start a bit earlier (think Ruby)... Educate maintainers to release proper .tar.gz, not braindead .gem packages containing the equivalent to an orig.tar.gz (but created due to a nice don't-ask-me-why-that's-not-properly-implemented bug in December 31, 1969)... And then complaining if you are distributing in stable anything older than their nightly checkouts. Yes, Perl and the CPAN rock my world, although my programming is nowadays mostly Ruby-based. The Ruby general mindset is WAY inferior. What? I admit I would have been able to parse the contents of your mail with the same success if you would have written in Spanish. :) Prehaps it is me who had to get up 4:20 this morning (so I started *really* early ;-) ) - but I do not even understand whether this is pro or contra proper watch files. Kind regards Andreas. -- http://fam-tille.de -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: On the subject of watchfiles (was Re: Proposed MBF: Debian upstream version higher than watch file-reported upstream version)
Andreas Tille dijo [Tue, Feb 26, 2008 at 03:02:31PM +0100]: > Well, in fact it is helpful if you teach upstream to organise releases > that way that watchfiles would work. This is not only in the interest > of Debian but for the whole FLOSS community so other interested users > will be able to transparantly download software as well and upstream > will start using a consistent version management. This will not work > for upstream dead software - but here are watch files void anyway. Heh, start a bit earlier (think Ruby)... Educate maintainers to release proper .tar.gz, not braindead .gem packages containing the equivalent to an orig.tar.gz (but created due to a nice don't-ask-me-why-that's-not-properly-implemented bug in December 31, 1969)... And then complaining if you are distributing in stable anything older than their nightly checkouts. Yes, Perl and the CPAN rock my world, although my programming is nowadays mostly Ruby-based. The Ruby general mindset is WAY inferior. -- Gunnar Wolf - [EMAIL PROTECTED] - (+52-55)5623-0154 / 1451-2244 PGP key 1024D/8BB527AF 2001-10-23 Fingerprint: 0C79 D2D1 2C4E 9CE4 5973 F800 D80E F35A 8BB5 27AF -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: On the subject of watchfiles (was Re: Proposed MBF: Debian upstream version higher than watch file-reported upstream version)
On Tue, 26 Feb 2008, Martín Ferrari wrote: Of course, we have luck, because CPAN (99% of our packages come from there, and we have only 4 unsolvable watch problems) is pretty well-behaving and consistent, compared to other upstreams. But chances are that watchfiles can be useful for the majority of people. Well, in fact it is helpful if you teach upstream to organise releases that way that watchfiles would work. This is not only in the interest of Debian but for the whole FLOSS community so other interested users will be able to transparantly download software as well and upstream will start using a consistent version management. This will not work for upstream dead software - but here are watch files void anyway. Kind regards Andreas, who had also doubt about watch files until about onw year ago ... -- http://fam-tille.de
Re: On the subject of watchfiles (was Re: Proposed MBF: Debian upstream version higher than watch file-reported upstream version)
Hi Jon, On Tue, Feb 26, 2008 at 11:20 AM, Jon Dowland <[EMAIL PROTECTED]> wrote: > > Is it worth investing much effort into debugging watch file > issues? In my experience, yes. I can cite the example of the perl group: 679 packages group maintained. You'll guess that there's no way in earth a small group of people can track such amount of packages manually. We heavily rely on the watchfiles for detecting new upstream releases, and the tool we use (http://pkg-perl.alioth.debian.org/cgi-bin/qareport.cgi) automates that for us. We go to great lengths to make every package have the correct watch information. Of course, we have luck, because CPAN (99% of our packages come from there, and we have only 4 unsolvable watch problems) is pretty well-behaving and consistent, compared to other upstreams. But chances are that watchfiles can be useful for the majority of people. -- Martín Ferrari
On the subject of watchfiles (was Re: Proposed MBF: Debian upstream version higher than watch file-reported upstream version)
Is it worth investing much effort into debugging watch file issues? In my experience, watchfiles are seldom useful. There was the whole problem with getting at HTTPS URLs; the sourceforge workarounds (that broke); etc. One of the packages I maintain (deutex) does not have the latest upstream version linked to from a website (it's referenced in a mailing list post somewhere). I've read several other examples of situations (unpredictable version number schemes etc.) where it falls short. Also, if a package is being looked after by an active maintenance team, you'd hope that they would be aware that a new upstream version was available: in many cases you'd hope they were aware one was *due*, often with pre-releases in experimental to catch issues for the larger suites. If a package is not being looked after by an active maintenance team, there's a bug in the package maintenance (or the package should be on it's way out); which won't be solved by tweaking the watch file. -- Jon Dowland -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]