Re: dehs and new watch files... was: Re: uscan over all packages....
- To: email@example.com
- Subject: Re: dehs and new watch files... was: Re: uscan over all packages....
- From: Jérôme Warnier <firstname.lastname@example.org>
- Date: Mon, 19 Jul 2004 23:23:47 +0200
- Message-id: <[🔎] email@example.com>
- In-reply-to: <Pine.LNX.firstname.lastname@example.org>
- References: <20040604145026.GA743@deb-support.de> <email@example.com> <20040608012701.GI25743@home.debsupport.de> <1086750169.11507.7.camel@atlantide> <20040609010549.GF13104@home.debsupport.de> <Pine.LNX.firstname.lastname@example.org>
Le mer 09/06/2004 à 09:27, Andreas Tille a écrit :
> On Wed, 9 Jun 2004, Michael Bramer wrote:
> > Because of this, I like to add/fix watch files to _all_ packages.
> In general I like the idea of working watch files for all packages.
> My personal experience showed in the past the following:
> 1. Upstream maintainers sometimes have a strange idea about numbering
> their source file versions, which makes it hard to write watch files.
The vast majority is predictable, though. And it is their interest to
provide with sensible numbering scheme, otherwise even users get
> 2. It is not always necessary to package each upstream version. Sometimes
> it make sense to stick to a certain version and wait until users
> ask explicitely (via wishlist "New version available" bug) for a
> new version. This keeps you in contact with your users on the
> one hand and others like to stabelize to certain versions.
It would still be the maintainer's decision to package or not, but he
would have a powerful tool to remind him, and even help him in many
> 3. Watch files do not work if packages are build from CVS sometimes.
> 4. Developers should be subscribed to the relevant mailing lists and
> so they should read the relevant announcements if upstream authors
> think there is something to announce and not when uscan finds a
> new version.
Both are not mutually exclusive.
> Well, this all were reasons against watch files - I just left out the pros
> because you know them. That's why I would love to take over watch files
> into my packages if someone volunteers to write them, but I personally see
> no relevance for me (personally) because I found for *my* packages that it
> is not worth in spending time to work around my reason named unter 1. above.
> > Comments?
> Well, just go for it and I'll accept and will fix bug reports with watchfile
> *patches* to my packages but I think I can spend my time more sanely (even if
> it is not much time).
I will take a look at that soon, but I don't know if you package
something I often use and find useful.
> Kind regards