Bug#793503: lintian: Please warn on obsolete URLs
Hi Jakub, Jakub Wilk wrote: > * Axel Beckert , 2015-09-02, 22:21: > ># Known obsolete websites / hosters who closed down or have frozen > ># content, one hostname per line. Subdomains will be matched, too. > > > >code.google.com > > $PROJECT.googlecode.com redirects to code.google.com/p/$PROJECT/, so > you should probably add "googlecode.com", too. Thanks for that hint! Will add it. Regards, Axel -- ,''`. | Axel Beckert , http://people.debian.org/~abe/ : :' : | Debian Developer, ftp.ch.debian.org Admin `. `' | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5 `-| 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
Bug#793503: lintian: Please warn on obsolete URLs
* Axel Beckert , 2015-09-02, 22:21: # Known obsolete websites / hosters who closed down or have frozen # content, one hostname per line. Subdomains will be matched, too. code.google.com $PROJECT.googlecode.com redirects to code.google.com/p/$PROJECT/, so you should probably add "googlecode.com", too. -- Jakub Wilk
Bug#793503: lintian: Please warn on obsolete URLs
On 2015-09-02 22:21, Axel Beckert wrote: > Hi, > > [...] > Niels: Which existing check would be best to add such a check? I'd > check at least debian/control, debian/copyright, debian/watch, and > debian/upstream/metadata (plus variants). Have you considered making a dedicated check for it? :) > Maybe fields.pm? The fields.pm check is testing fields in the .dsc file. Not to confused with control-file.pm which tests the d/control file (which itself should not to be confused with control-file*s*.pm. The latter checks the control files in binary packages). > debian/copyright doesn't necessarily need to have fields (i.e. if not > in DEP5 format) and debian/watch doesn't have fields either... > All the more reason for it to not go into fields.pm. The other being that fields.pm is rather large already. > Maybe split up the test using the same data file (I'd name it > data/*/obsolete-sites) and then using it in fields.pm, watch-file.pm > and copyright-file.pm separately? That is certainly an option. In that case, please put the data file in data/common. Please keep in mind that the Lintian::Data->new call should be *exactly the same* in all checks using it. A bit of a caveat, but in practise it has not been an issue to date. > But where to check debian/upstream/metadata and friends then? > > [...] > > Regards, Axel > I do not think we look at the file at all. You would have to make such a check. :) Thanks, ~Niels signature.asc Description: OpenPGP digital signature
Bug#793503: lintian: Please warn on obsolete URLs
Hi, Guillem Jover wrote: > On Mon, 2015-08-24 at 11:00:18 +0300, Riku Voipio wrote: > > These obsolete urls are already checked with duck[1][2]. I > > think what would make sense would be to make lintian recommend > > duck. Then lintian can run duck if it has been installed. > > I was aware of duck, but as stated by Jakub, that is a non-option for > lintian. In addition many of these URLs are (or were at the time) > still reachable, but might disappear in the near future, and we know > beforehand that they are obsolete, so I think they are really good > candidates for a lintian check, because the just require a string > match instead of a network check. Yep, one more for which I filed a bug report (#797815) today: freshmeat.net / freecode.com I'll think about such a test for working but known to be legacy URLs. What about the tag name "legacy-url-in-packaging"? Or rather "obsolete-url-in-packaging"? Niels: Which existing check would be best to add such a check? I'd check at least debian/control, debian/copyright, debian/watch, and debian/upstream/metadata (plus variants). Maybe fields.pm? debian/copyright doesn't necessarily need to have fields (i.e. if not in DEP5 format) and debian/watch doesn't have fields either... Maybe split up the test using the same data file (I'd name it data/*/obsolete-sites) and then using it in fields.pm, watch-file.pm and copyright-file.pm separately? But where to check debian/upstream/metadata and friends then? So far my data file looks like this: ---8<--- # Known obsolete websites / hosters who closed down or have frozen # content, one hostname per line. Subdomains will be matched, too. code.google.com gitorious.org codehaus.org freshmeat.net freecode.com --->8--- Regards, Axel -- ,''`. | Axel Beckert , http://people.debian.org/~abe/ : :' : | Debian Developer, ftp.ch.debian.org Admin `. `' | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5 `-| 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
Bug#793503: lintian: Please warn on obsolete URLs
Hi Riku, Jakub Wilk wrote: > * Riku Voipio , 2015-08-24, 11:00: > >These obsolete urls are already checked with duck[1][2]. I think > >what would make sense would be to make lintian recommend duck. > >Then lintian can run duck if it has been installed. > > Leaving aside privacy issues, that would make Lintian output > dependent on external world, which would be against its design > constraints: > https://lintian.debian.org/manual/section-1.3.html > > So I'm afraid we can't run duck, at least not by default. I think a nice place to have duck being run automatically via some per-user configuration is the debuild script from devscripts. It runs lintian by default after each build and has a hook for running further such scripts. Unfortunately this hook is not run inside the source directory but in its parent. So it needs a bit of shell code around it. And debuild exits if the final error code is not zero. So it also needs "true" as last command. I currently use this: DEBUILD_LINTIAN_HOOK="if [ -d %p-%u ]; then cd %p-%u; duck; elif [ -d %p ]; then cd %p; duck; else exit 1; fi; true" (Simon: Maybe this is suitable as hint for the duck man page?) Another way is to use wrapper tools like these: https://anonscm.debian.org/cgit/pkg-perl/packages/pkg-perl-tools.git/tree/examples/check-build https://packages.debian.org/source/experimental/check-all-the-things Regards, Axel -- ,''`. | Axel Beckert , http://people.debian.org/~abe/ : :' : | Debian Developer, ftp.ch.debian.org Admin `. `' | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5 `-| 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
Bug#793503: lintian: Please warn on obsolete URLs
Hi! On Mon, 2015-08-24 at 11:00:18 +0300, Riku Voipio wrote: > These obsolete urls are already checked with duck[1][2]. I > think what would make sense would be to make lintian recommend > duck. Then lintian can run duck if it has been installed. I was aware of duck, but as stated by Jakub, that is a non-option for lintian. In addition many of these URLs are (or were at the time) still reachable, but might disappear in the near future, and we know beforehand that they are obsolete, so I think they are really good candidates for a lintian check, because the just require a string match instead of a network check. Thanks, Guillem
Bug#793503: lintian: Please warn on obsolete URLs
-BEGIN PGP SIGNED MESSAGE- Hash: SHA512 Am 2015-08-25 um 16:20 schrieb Jakub Wilk: > Hi Riku! > > * Riku Voipio , 2015-08-24, 11:00: >> These obsolete urls are already checked with duck[1][2]. I think >> what would make sense would be to make lintian recommend duck. >> Then lintian can run duck if it has been installed. > > Leaving aside privacy issues, that would make Lintian output > dependent on external world, which would be against its design > constraints: https://lintian.debian.org/manual/section-1.3.html > > So I'm afraid we can't run duck, at least not by default. > I fully agree with that, my original intent (before i made duck as a seperate tool) was to integrate URL checks into lintian as new check modules. But this is currently a no-go. Because there are currently way more projects with e.g. code.google.com based URLS than there are packages with issues (e.g. 404, no repo exported,...) reported by duck, i will extend the duck website to list packages grouped by domain and post this on debian-devel . Especially for code.google.com this is somewhat urgent, as of [1], excer t: " March 12, 2015 - New project creation disabled. August 24, 2015 - The site goes read-only. You can still checkout/view project source, issues, and wikis. January 25, 2016 - The project hosting service is closed. You will be able to download a tarball of project source, issues, and wikis. These tarballs will be available throughout the rest of 2016. " This would eventually raise some attention and help to prevent bitrod. Simon [1] http://google-opensource.blogspot.fi/2015/03/farewell-to-google-code.htm l -BEGIN PGP SIGNATURE- Version: GnuPG v2 iQIcBAEBCgAGBQJV3Wm/AAoJEBy08PeN7K/pWMgP/1BKWevj2hUuspUDWftOM2Ts 3Abl57xeI65XPuaykN1Ix5nTTwBEyOxy2tmAfW48l2K5W+r414V1GGmAzopAkDhd SBZobaVzkT2wSCXtV1IrmTYmPLaHltVvbvUrhzxsg8NVgwe2TR+Lf25b6wFZs4nf VWabmFTHjf/n9Ceh1e/SRa1Tme8GBxKDDvqy1jcLvUb1/8ODvfLYgBrdzAVyxLFR 5NYfw5ZVfuMqV0edjtx/GSK1SsxsGSAG1fJ5YPm2GKLklXfMjOxG6pWdIUbordvM BSJLj5QaC7cYViJrv4t9P4VN275P9QJIwFqxtkwJrxhBhQRMV1wml7K1W9X3oD0k v47r0MWgi8QippxvzY/5gaL678OGQvtocvsuRzNKgFOkGTcKFer976qgPQThimLi jfxLSeXAZbIKyFimCAZN3YVEOL/1x/8P0L/FjkVXm2lB80ilbIu1FzWcewaLRPA7 AI+xN08+I/nD7cpZ/OA1iOBFDyOFWn4r7/fzFzmkZdnXqWtTJQETDOO4iRhDbeQE FVXUEKHDTOSFKA+fGHQiul/frfHm/JvexsT6gmQw9yl+DLFjGNSb2WD+nGZ06zxE n6sQbJ5HTN0U/bzQjHCR35ZNvAQeUF0Ati52uBbxB3ROKVVX5NNY+tscFbk0GD3j SVqQ0CI9r5rUvUj0Zt/C =hQx4 -END PGP SIGNATURE-
Bug#793503: lintian: Please warn on obsolete URLs
Hi Riku! * Riku Voipio , 2015-08-24, 11:00: These obsolete urls are already checked with duck[1][2]. I think what would make sense would be to make lintian recommend duck. Then lintian can run duck if it has been installed. Leaving aside privacy issues, that would make Lintian output dependent on external world, which would be against its design constraints: https://lintian.debian.org/manual/section-1.3.html So I'm afraid we can't run duck, at least not by default. -- Jakub Wilk
Bug#793503: lintian: Please warn on obsolete URLs
Hi, These obsolete urls are already checked with duck[1][2]. I think what would make sense would be to make lintian recommend duck. Then lintian can run duck if it has been installed. [1] http://duck.debian.net/ [2] https://packages.debian.org/sid/duck
Bug#793503: lintian: Please warn on obsolete URLs
Package: lintian Version: 2.5.35 Severity: wishlist Hi! There are several URLs that are known to be obsolete, and that even if they currently work, will disappear in the future or point to stale sites. Among those there are at least: code.google.com gitorious.org codehaus.org I think it would make sense to warn on those whenever they appear at least in any of the Vcs or Homepage fields. Thanks, Guillem -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
