Bug#808538: RFS: corebird
Hi, On Wed, Dec 30, 2015 at 11:47:16AM +0800, Paul Wise wrote: > If the upstream build system supports parallel builds, you are just > wasting buildd time by removing that option, please put it back in. > > The only reason that it isn't the default yet is because not all > upstream build systems work when run in parallel mode. Ok cool. It's not worth doing another upload to add that, but it can be added in on the next upload of the package. When I asked before about this, although this was a while ago, I was told the buildds would only use one core per build anyway so I didn't think this would be useful. I guess this has changed now. Thanks, Iain. --
Bug#808538: RFS: corebird
Hi Iain, thanks for the detailed review! On Thu, 24 Dec 2015 21:37:31 + "Iain R. Learmonth"wrote: > Hi, > > Unfortunately this is not yet ready for upload. It's nearly there though. Good to hear :). > debian/copyright: This is one of the most important things to get right, > and you have got this right. Do be careful though licensing your packaging > under different terms to the upstream code. In this case, it's possible to > take patches you produce and relicense them under GNU GPL v3 (due to the > "any later version" clause) so they can be integrated upstream, but this > would not be the case with incompatible licenses. I prefer to always just > use the license that upstream used. Hm, good point, I just use GPL-2+ for all my debian/* files. I'll think about it but keep it for now. > debian/control: You've set the section to "misc". I would suggest you set > this to "web" instead. See Debian Policy §2.4 for more detail on sections. done. > The extended description also needs to be slightly more extended. Please > talk about the features this application has. done. > debian/rules: You've got some things in here that I don't believe are > necessary. I like the debian/rules to be as clean as possible, using as many > of the current defaults as possible. The following statements are redundant > as you are using compat level 9: > > export DEB_BUILD_MAINT_OPTIONS = hardening=+all > export DEB_LDFLAGS_MAINT_APPEND = -Wl,--as-needed I also like to use defaults but here, hardening of builds seems to be important enough to deviate from defaults. > It is acceptable to leave in the "nocheck" export as long as that's not a > permanent thing. Please contact upstream and ask them to make the tests > pass. I'll contact upstream and try to solve this. > export V = 1 > > I'm not sure what $V does. Is this really necessary for the release in the > Debian archive? As Mattia explained it produces more verbose build logs which helped me to understand failing builds on buildds in the past. > You have an "override_dh_auto_configure" section, please look up > dh-autoreconf as this will handle this for you. Currently this package is > not policy compliant as "debian/rules clean" will not restore the package to > its original state. dh-autoreconf handles storing the old versions for you > and ensuring the package is policy compliant in that respect. Thanks for catching this, I changed it now. > dh $@ --parallel > > Unless you have a good reason for the --parallel, please remove this. If it > turns out later that the Debian build systems prefer this option, it would > become a default. As I said above, I prefer defaults be used whenever > possible. I removed the "--parallel" now. > The package bulds fine using cowbuilder, and lintian has no futher > complaints, if you can fix these issues then I will be happy to sponsor this > package for you. I hope I addressed all your suggestions. I uploaded a revised build to http://mentors.debian.net/package/corebird Best, Philip signature.asc Description: OpenPGP digital signature
Bug#808538: RFS: corebird
On Wed, Dec 30, 2015 at 2:16 AM, Philip Rinn wrote: > I removed the "--parallel" now. If the upstream build system supports parallel builds, you are just wasting buildd time by removing that option, please put it back in. The only reason that it isn't the default yet is because not all upstream build systems work when run in parallel mode. -- bye, pabs https://wiki.debian.org/PaulWise
Bug#808538: RFS: corebird
Hi, On Thu, Dec 24, 2015 at 09:37:31PM +, Iain R. Learmonth wrote: > export DEB_BUILD_MAINT_OPTIONS = hardening=+all > export DEB_LDFLAGS_MAINT_APPEND = -Wl,--as-needed Apologies, I was wrong. These are not enabled by default with compat level 9. As these options may be useful for some packages, I will leave it to you to decide whether or not you still want these here. Thanks, Iain. --
Bug#808538: RFS: corebird/1.1-1 [ITP]
owner 808538 ! kthxbye Hi, Just taking a look at this now. Thanks, Iain. --
Bug#808538: RFS: corebird
Hi, Unfortunately this is not yet ready for upload. It's nearly there though. debian/copyright: This is one of the most important things to get right, and you have got this right. Do be careful though licensing your packaging under different terms to the upstream code. In this case, it's possible to take patches you produce and relicense them under GNU GPL v3 (due to the "any later version" clause) so they can be integrated upstream, but this would not be the case with incompatible licenses. I prefer to always just use the license that upstream used. debian/control: You've set the section to "misc". I would suggest you set this to "web" instead. See Debian Policy §2.4 for more detail on sections. The extended description also needs to be slightly more extended. Please talk about the features this application has. debian/rules: You've got some things in here that I don't believe are necessary. I like the debian/rules to be as clean as possible, using as many of the current defaults as possible. The following statements are redundant as you are using compat level 9: export DEB_BUILD_MAINT_OPTIONS = hardening=+all export DEB_LDFLAGS_MAINT_APPEND = -Wl,--as-needed It is acceptable to leave in the "nocheck" export as long as that's not a permanent thing. Please contact upstream and ask them to make the tests pass. export V = 1 I'm not sure what $V does. Is this really necessary for the release in the Debian archive? You have an "override_dh_auto_configure" section, please look up dh-autoreconf as this will handle this for you. Currently this package is not policy compliant as "debian/rules clean" will not restore the package to its original state. dh-autoreconf handles storing the old versions for you and ensuring the package is policy compliant in that respect. dh $@ --parallel Unless you have a good reason for the --parallel, please remove this. If it turns out later that the Debian build systems prefer this option, it would become a default. As I said above, I prefer defaults be used whenever possible. The package bulds fine using cowbuilder, and lintian has no futher complaints, if you can fix these issues then I will be happy to sponsor this package for you. Thanks, Iain. --
Bug#808538: RFS: corebird
disclaimer: I did not look at the package at all. On Thu, Dec 24, 2015 at 09:37:31PM +, Iain R. Learmonth wrote: > export V = 1 > > I'm not sure what $V does. Is this really necessary for the release in the > Debian archive? Some people (me, for instance) like to have the build as verbose as possible. locally I have DH_VERBOSE exported in my build env. For this debhelper started passing a --disable-silent-rules to autoconf and setting something else for other build systems, for example. I'd love to see such common variables exported too, and ISTR there is a bug for something like this. > dh $@ --parallel > > Unless you have a good reason for the --parallel, please remove this. If it > turns out later that the Debian build systems prefer this option, it would > become a default. This is going to be the default for debhelper compat 10. I tend to force --parallel for my packages when the build takes more than 10 minutes otherwise and in my build env I have DEB_BUILD_OPTIONS=parallel=4, since that way I have a huge performance boost. Also this is nice for buildds (not only buildd.d.o, I'm also talking about the ones rebuilding the archive, like rb.d.n, where the throughput is kinda important), so they can waste less time on your package and use better their CPU. > As I said above, I prefer defaults be used whenever possible. Your call :) I'm just precising some bits. This kind of stuff is really in the YMMV / YKINMKBYKIOK field :D -- regards, Mattia Rizzolo GPG Key: 66AE 2B4A FCCF 3F52 DA18 4D18 4B04 3FCD B944 4540 .''`. more about me: http://mapreri.org : :' : Launchpad user: https://launchpad.net/~mapreri `. `'` Debian QA page: https://qa.debian.org/developer.php?login=mattia `- signature.asc Description: PGP signature
Bug#808538: RFS: corebird/1.1-1 [ITP]
Package: sponsorship-requests Severity: wishlist Dear mentors, I am looking for a sponsor for my package "corebird" Package name: corebird Version : 1.1-1 Upstream Author : Timm BäderURL : http://corebird.baedert.org License : GPL-3+ Section : misc It builds those binary packages: corebird - Native Gtk+ Twitter client for the Linux desktop To access further information about this package, please visit the following URL: http://mentors.debian.net/package/corebird Alternatively, one can download the package with dget using this command: dget -x http://mentors.debian.net/debian/pool/main/c/corebird/corebird_1.1-1.dsc More information about corebird can be obtained from http://corebird.baedert.org. Regards, Philip Rinn -- System Information: Debian Release: stretch/sid APT prefers testing APT policy: (600, 'testing'), (550, 'unstable'), (500, 'testing-updates') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 4.2.0-1-amd64 (SMP w/4 CPU cores) Locale: LANG=de_DE.utf8, LC_CTYPE=de_DE.utf8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Init: systemd (via /run/systemd/system)