Re: missing input and more

2015-02-05 Thread 宋文武
2015-02-05 15:05 GMT+08:00 : > Hi Guix, > > during my packaging attempts of connman (which looks good so far) I > found that gnutls is missing "install-time" dependency - zlib. > > 1] ldd shows run-time dependency: > $ ldd > /gnu/store/0mfgwxgqyhlvi5xyi1j2lbr3ygczsdyp-gnutls-3.2.21/lib/libgnutls.s

Re: Helping programs find their icons

2015-02-05 Thread Andreas Enge
Would it not help to add hicolor-icon-theme as a propagated input to some packages? Or maybe add it as a propagated input inside glib-or-gtk-build-system if that is possible? Andreas

Re: [PATCH] gnu: gnutls: Configure location of system-wide trust store

2015-02-05 Thread Andreas Enge
Hello Mark, On Wed, Feb 04, 2015 at 10:35:57AM -0500, Mark H Weaver wrote: > Since GnuTLS updates require so much rebuilding, I would prefer to wait > for Hydra to finish the rebuild of intel platforms on another branch > before merging into master. I just cherry-picked the new gnutls into master

Re: missing input and more

2015-02-05 Thread Andreas Enge
On Thu, Feb 05, 2015 at 08:05:03AM +0100, sleep_wal...@suse.cz wrote: > 1] ldd shows run-time dependency: > $ ldd > /gnu/store/0mfgwxgqyhlvi5xyi1j2lbr3ygczsdyp-gnutls-3.2.21/lib/libgnutls.so | > grep libz >libz.so.1 => > /gnu/store/x8cg3irwf8y2mkr88bqmsqhadi71xf6s-zlib-1.2.7/lib/libz.so.

issues with offloading

2015-02-05 Thread Ricardo Wurmus
Hi Guix, I'm trying to set up offloading such that I can have a powerful build host compile stuff for me when my workstation isn't strong enough. I bumped into a couple of issues while doing this, prompting me to write this email to discuss what might be changed to improve this. * lsh required

Re: missing input and more

2015-02-05 Thread 宋文武
Andreas Enge writes: > On Thu, Feb 05, 2015 at 08:05:03AM +0100, sleep_wal...@suse.cz wrote: >> 1] ldd shows run-time dependency: >> $ ldd >> /gnu/store/0mfgwxgqyhlvi5xyi1j2lbr3ygczsdyp-gnutls-3.2.21/lib/libgnutls.so | >> grep libz >>libz.so.1 => >> /gnu/store/x8cg3irwf8y2mkr88bqmsqhad

New Esperanto PO file for 'guix-packages' (version 0.8.1)

2015-02-05 Thread Translation Project Robot
Hello, gentle maintainer. This is a message from the Translation Project robot. A revised PO file for textual domain 'guix-packages' has been submitted by the Esperanto team of translators. The file is available at: http://translationproject.org/latest/guix-packages/eo.po (We can arrange t

Re: [PATCH] gnu: Add ngircd.

2015-02-05 Thread Taylan Ulrich Bayırlı/Kammer
I had time to tinker around and, with willingness from the ngIRCd developers to patch their test suite, came up with a relatively simple solution. The relevant patch to the test suite is not necessarily as portable as ngIRCd requires it to be (or maybe it is) so this might not be the patch that wi

Re: missing input and more

2015-02-05 Thread Tomas Cech
At Thu, 05 Feb 2015 19:01:18 +0800, 宋文武 wrote: > > Andreas Enge writes: > > > On Thu, Feb 05, 2015 at 08:05:03AM +0100, sleep_wal...@suse.cz wrote: > >> 1] ldd shows run-time dependency: > >> $ ldd > >> /gnu/store/0mfgwxgqyhlvi5xyi1j2lbr3ygczsdyp-gnutls-3.2.21/lib/libgnutls.so > >> | grep libz

Re: [PATCH]: Add Ant.

2015-02-05 Thread Ludovic Courtès
Ricardo Wurmus skribis: > this patch adds ant-minimal, a minimal configuration of Apache Ant. Ant > can be configured with lots of optional libraries, but since most of > these libraries are built with Ant, I think it makes sense to provide a > minimal version. Yes. > A wart is that Ant (even

Re: [PATCH]: Add Ant.

2015-02-05 Thread Ludovic Courtès
Ricardo Wurmus skribis: > A couple more things: > > * the installation script copies MS DOS batch scripts to $out/bin, which > we could do without. Indeed. :-) Would be best to get rid of them. > * there are many launcher scripts for Ant: bash scripts, perl scripts, > Python scripts ... D

Re: problem with building gcc-cross-4.8.3 for i686-pc-gnu

2015-02-05 Thread Ludovic Courtès
Marek Benc skribis: > The attached patch does the following: > 1.) Adds the name of the dynamic linker to gnu/packages/bootstrap.scm > 2.) Adds code to gnu/package/gcc to substitute the default dynamic > linker in gcc/config/arch/gnu.h for the one to be used. > 3.) Changes how the build sys

Re: Helping programs find their icons

2015-02-05 Thread 宋文武
Mark H Weaver writes: > Hello Guix, > > I've investigated why wicd-gtk is unable to find its icons and fixed it. > Looking at the output of strace, I found that when wicd-gtk looks for > icons, it searches for 'index.theme' files in many different directories > of the form: > > /gnu/store/.../s

Re: problem with building gcc-cross-4.8.3 for i686-pc-gnu

2015-02-05 Thread Ludovic Courtès
Marek Benc skribis: > On 01/31/2015 11:13 PM, Marek Benc wrote: > >> >> The current issue is that, when building cross-gcc, when it reaches >> libgomp, configure fails when trying to create a dummy binary. For some >> reason, the compiled cross-compiler needs --rpath to specify where >> shared li

Re: [PATCH] gnu: Add ngircd.

2015-02-05 Thread Ludovic Courtès
taylanbayi...@gmail.com (Taylan Ulrich "Bayırlı/Kammer") skribis: > From 34b2efd8244caf0cd16c1aa510209a54def148b7 Mon Sep 17 00:00:00 2001 > From: =?UTF-8?q?Taylan=20Ulrich=20Bay=C4=B1rl=C4=B1/Kammer?= > > Date: Sat, 31 Jan 2015 22:03:25 +0100 > Subject: [PATCH] gnu: Add ngircd. > > * gnu/packag

Orc fails to build

2015-02-05 Thread Andreas Enge
Hello, orc fails its tests on i686 and mips on hydra: http://hydra.gnu.org/build/233541 http://hydra.gnu.org/build/233919/nixlog/1/tail-reload , together with gstreamer* for which it is an input. The failing function is test-limits. Something to do with 32/64 bit? Andreas

Re: Guix System Distribution now FSF-endorsed

2015-02-05 Thread Daniel Pimentel
On 2015-02-04 19:19, l...@gnu.org wrote: "jaccas (pjfcl)" skribis: On the distribution have been included in the FSF list and for the chosen name. Simple is better. Thanks! I actually forgot to mention it on this list: http://www.fsf.org/news/fsf-adds-guix-system-distribution-to-list-of-e

Re: [PATCHES] Add wicd and wicd-service

2015-02-05 Thread Daniel Pimentel
On 2015-02-04 22:22, Mark H Weaver wrote: Mark H Weaver writes: * Unless run as root, the user interfaces to configure wicd via dbus always fail. The wicd dbus configuration file is supposed to allow anyone in the 'netdev' group to access and configure the daemon, but this doesn't work.

Re: Orc fails to build

2015-02-05 Thread 宋文武
Andreas Enge writes: > Hello, > > orc fails its tests on i686 and mips on hydra: >http://hydra.gnu.org/build/233541 >http://hydra.gnu.org/build/233919/nixlog/1/tail-reload , > together with gstreamer* for which it is an input. > > The failing function is test-limits. Something to do with

Re: Guix System Distribution now FSF-endorsed

2015-02-05 Thread Taylan Ulrich Bayırlı/Kammer
Daniel Pimentel writes: > In soon the Guix GSD will be in distrowatch too, I request it. This makes me worry because DistroWatch seems strongly focused on "Linux distros". Maybe it will be fine if the description states clearly what Guix System Distribution really is... (I'm aware that Trisque

Re: [PATCH] gnu: Add ngircd.

2015-02-05 Thread Taylan Ulrich Bayırlı/Kammer
l...@gnu.org (Ludovic Courtès) writes: > taylanbayi...@gmail.com (Taylan Ulrich "Bayırlı/Kammer") skribis: > >> From 34b2efd8244caf0cd16c1aa510209a54def148b7 Mon Sep 17 00:00:00 2001 >> From: =?UTF-8?q?Taylan=20Ulrich=20Bay=C4=B1rl=C4=B1/Kammer?= >> >> Date: Sat, 31 Jan 2015 22:03:25 +0100 >> Su

Re: Guix System Distribution now FSF-endorsed

2015-02-05 Thread Ricardo Wurmus
Daniel Pimentel writes: > In soon the Guix GSD will be in distrowatch too, I request it. The "Guix Guix System Distribution", eh? ;) When using the abbreviation "GSD" for "Guix System Distribution" I recommend prefixing it with "GNU" to avoid confusion. ~~ Ricardo

Re: Guix System Distribution now FSF-endorsed

2015-02-05 Thread Thompson, David
On Thu, Feb 5, 2015 at 9:33 AM, Ricardo Wurmus wrote: > > Daniel Pimentel writes: >> In soon the Guix GSD will be in distrowatch too, I request it. > > The "Guix Guix System Distribution", eh? ;) > > When using the abbreviation "GSD" for "Guix System Distribution" I > recommend prefixing it with "

Re: Guix System Distribution now FSF-endorsed

2015-02-05 Thread Daniel Pimentel
On 2015-02-05 11:28, taylanbayi...@gmail.com wrote: Daniel Pimentel writes: In soon the Guix GSD will be in distrowatch too, I request it. This makes me worry because DistroWatch seems strongly focused on "Linux distros". Maybe it will be fine if the description states clearly what Guix S

Re: Guix System Distribution now FSF-endorsed

2015-02-05 Thread Daniel Pimentel
On 2015-02-05 11:33, Ricardo Wurmus wrote: Daniel Pimentel writes: In soon the Guix GSD will be in distrowatch too, I request it. The "Guix Guix System Distribution", eh? ;) When using the abbreviation "GSD" for "Guix System Distribution" I recommend prefixing it with "GNU" to avoid confusion

Re: Orc fails to build

2015-02-05 Thread Andreas Enge
On Thu, Feb 05, 2015 at 10:18:48PM +0800, 宋文武 wrote: > Find it: > https://bugzilla.gnome.org/show_bug.cgi?id=735273 > So I think we have to disable the tests? Good find! If you can without too much effort, it would be nice to add a patch that disables exactly this test. In any case, please add a

Re: [PATCH] gnu: Add ngircd.

2015-02-05 Thread Mark H Weaver
taylanbayi...@gmail.com (Taylan Ulrich "Bayırlı/Kammer") writes: > +(arguments > + `(#:configure-flags > + `("--with-gnutls" "--with-iconv" "--enable-ipv6" > + ,,@(if (string-suffix? "-linux" > +(or (%current-target-system) > +

Re: missing input and more

2015-02-05 Thread Andreas Enge
On Thu, Feb 05, 2015 at 07:01:18PM +0800, 宋文武 wrote: > Ah, I think libraries in 'Requires.private' should also be propagated, > otherwise, we got: > $ pkg-config --cflags gnutls > = Package 'zlib', required by 'GnuTLS', not found > If not propagated, other packages use gnutls have to add zlib t

Re: missing input and more

2015-02-05 Thread Andreas Enge
On Thu, Feb 05, 2015 at 12:57:27PM +0100, Tomas Cech wrote: > But that fixed only one problem while we should do something dependencies in > general. > > Imagine I have not came with this bug and some other people would just added > zlib to the inputs. We could have it wrong on multiple places and

[PATCH 0/4] gnu: Fix ld.so detection of cross-compilers for the GNU Hurd system.

2015-02-05 Thread Marek Benc
On 02/05/2015 01:38 PM, Ludovic Courtès wrote: This all looks good to me. However, this should really be 5 different patches so as to ease review, bisection, etc. Could you do that? Please prefer one message per patch, as with ‘git send-email’. Sure thing, just finished splitting/converti

[PATCH 2/4] gnu: cross-base: Make it aware of non-Linux (ie. Hurd)

2015-02-05 Thread Marek Benc
This fixes the problem Manolis was facing earlier, with string-append getting an #f argument. -- Marek. >From 4dc55bfe59bf60c4f55a7e54da2f5ee17589ff9e Mon Sep 17 00:00:00 2001 From: Marek Benc Date: Thu, 5 Feb 2015 17:01:14 +0100 Subject: [PATCH] gnu: cross-base: Make it aware of non-Linux (ie.

[PATCH 4/4] gnu: gcc: Also substitute the dynamic linker name for GNU, (ie. Hurd) systems.

2015-02-05 Thread Marek Benc
In the gcc package, the system-dependent paths to the dynamic linker get substituted before building, but only for Linux systems. This patch adds the substitution for GNU systems. -- Marek. >From 7ae53fd611dd78a8c185734dad268428a8b931a8 Mon Sep 17 00:00:00 2001 From: Marek Benc Date: Thu, 5 Fe

[PATCH 1/4] gnu: Move the cross-gcc build system arguments into a separate procedure.

2015-02-05 Thread Marek Benc
This one is actually just a copy/paste from the master branch. It includes a fix to the problem of cross compilers compiling binaries with the building system's ld.so in mind, I take no credit for this. -- Marek. >From b4665f3e98d70e6efb34b0f945477b4336a51e7d Mon Sep 17 00:00:00 2001 From: Mare

[PATCH 3/4] gnu: bootstrap: Add the location of where ld.so is located on Hurd systems.

2015-02-05 Thread Marek Benc
This one should be pretty self explanatory. -- Marek. >From 098ca112becb95199d3e66078b75fbf5b9b54a69 Mon Sep 17 00:00:00 2001 From: Marek Benc Date: Thu, 5 Feb 2015 17:07:18 +0100 Subject: [PATCH] gnu: bootstrap: Add the location of where ld.so is located on Hurd systems. * gnu/packages/bootst

Re: [PATCH 1/4] gnu: Move the cross-gcc build system arguments into a separate procedure.

2015-02-05 Thread Mark H Weaver
Marek Benc writes: > This one is actually just a copy/paste from the master branch. It > includes a fix to the problem of cross compilers compiling binaries > with the building system's ld.so in mind, I take no credit for this. Please don't copy/paste from the master branch. Merging the master b

Re: [PATCH 1/4] gnu: Move the cross-gcc build system arguments into a separate procedure.

2015-02-05 Thread Mark H Weaver
Mark H Weaver writes: > Marek Benc writes: >> This one is actually just a copy/paste from the master branch. It >> includes a fix to the problem of cross compilers compiling binaries >> with the building system's ld.so in mind, I take no credit for this. > > Please don't copy/paste from the mast

wip-hurd branch rebased with master

2015-02-05 Thread Manolis Ragkousis
These patches are the commits from wip-hurd branch rebased on top of master. From e7db980d40dec4623488e44be4304ec54c2ae354 Mon Sep 17 00:00:00 2001 From: Manolis Ragkousis Date: Tue, 27 Jan 2015 18:39:47 + Subject: [PATCH 2/2] gnu: cross-base: Add support to cross-build libc for GNU/Hurd. MI

Re: wip-hurd branch rebased with master

2015-02-05 Thread Manolis Ragkousis
I was saying to Ludovic at fosdem that when rebasing to master, trying to build glibc/hurd you get the error: In unknown file: ?: 0 [string-append "--with-headers=" #f "/include"] ERROR: In procedure string-append: ERROR: In procedure string-append: Wrong type (expecting string): #f the error

Re: [PATCH 3/4] gnu: bootstrap: Add the location of where ld.so is located on Hurd systems.

2015-02-05 Thread Manolis Ragkousis
On 5 February 2015 at 16:29, Marek Benc wrote: > This one should be pretty self explanatory. I have done the same on my local branch so this one can be safely applied on wip-hurd branch.

Re: missing input and more

2015-02-05 Thread Ludovic Courtès
Andreas Enge skribis: > Maybe we could extend "guix lint" so that it compiles the package and performs > checks on the output (for instance, looking for .pc files and checking > whether libraries are correctly propagated)? That would be costly and probably difficult to automate: it’s almost doab

Re: problem with building gcc-cross-4.8.3 for i686-pc-gnu

2015-02-05 Thread Marek Benc
On 02/05/2015 09:26 PM, Ludovic Courtès wrote: Marek Benc skribis: On 02/05/2015 01:44 PM, Ludovic Courtès wrote: Another thing I don’t understand is that we use "--disable-libgomp" in ‘cross-gcc-arguments’, which means that the first cross-gcc (sans libc) does not build libgomp. What ar

Re: Helping programs find their icons

2015-02-05 Thread Ludovic Courtès
Mark H Weaver skribis: > However, this problem is quite widespread. Many of our gtk packages > install icons into $out/share/icons/hicolor but have no 'index.theme', > and I suspect that's the reason the xfce application menu is missing so > many icons. > > inkscape-0.91 > emacs-24.4 > gtk

Re: 01/01: gnu: gnutls: Propagate zlib.

2015-02-05 Thread Ludovic Courtès
??? skribis: > commit 122ddc595e22a933dbffaad1153876cc7ea2203b > Author: 宋文武 > Date: Thu Feb 5 18:46:59 2015 +0800 > > gnu: gnutls: Propagate zlib. > > * gnu/packages/gnutls.scm (gnutls): Move zlib to 'propagated-inputs'. [...] > (inputs > `(("guile" ,guile-2.0) > -

Re: [PATCH] Add 'waf-build-system'.

2015-02-05 Thread Ricardo Wurmus
>>> From adc52a74fb12943fd77c97bf75a2092e839f9024 Mon Sep 17 00:00:00 2001 >>> From: Ricardo Wurmus >>> Date: Wed, 28 Jan 2015 22:39:03 +0100 >>> Subject: [PATCH] Add 'waf-build-system'. >>> >>> * guix/build-system/waf.scm, guix/build/waf-build-system.scm: New files. >>> * Makefile.am (MODULES):

Re: issues with offloading

2015-02-05 Thread Ludovic Courtès
Ricardo Wurmus skribis: > * lsh required > > The manual does not appear to mention that for offloading lsh is > expected to be installed on the submitting host. Since I only had > OpenSSH installed (on the local workstation and the remote server) I > decided to redefine %lsh-command and