public/nixpkgs
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
136,354 Commits 1 Branch 0 Tags
77f9ebc8ada5a1aabcedeaadab70dab673255dbf
Commit Graph
100 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Will Dietz c95720ca14 procps-ng: 3.3.12 -> 3.3.13 
https://gitlab.com/procps-ng/procps/tags/v3.3.13

Patch to fix new version w/musl.
 2018-04-05 12:40:27 -05:00
Will Dietz ce4a423499 busybox: 1.28.2 -> 1.28.3 2018-04-05 12:00:51 -05:00
Will Dietz bf44cf71b7 mesa: 17.3.7 -> 17.3.8 
https://www.mesa3d.org/relnotes/17.3.8.html

Remove upstreamed portion of musl patch,
always add remaining fix (missing include).
 2018-04-05 11:41:49 -05:00
Will Dietz 3bdcd3e298 mpg123: 1.25.8 -> 1.25.10 
1.25.10: https://www.mpg123.de/cgi-bin/news.cgi#2018-03-05
1.25.9 was replaced by 1.25.10 before it was announced (see link above).
 2018-04-03 14:22:27 -05:00
Will Dietz 0086036141 Machine.pm: tweak to look for graphical target in waitForX 
Current message may not be printed
(nothing in systemd contains that string,
 it comes from elsewhere)
 2018-04-03 13:59:16 -05:00
Will Dietz 1d16384bb6 mesa: use llvm6 2018-04-03 08:11:27 -05:00
Will Dietz 231cdaf159 mesa: 17.3.6 -> 17.3.7 2018-04-03 08:11:26 -05:00
Will Dietz 4cee817a31 gnumake42HEAD: remove attribute to fix eval 
This was removed as part of
519f0b8db2
but is still present for whatever reason.

Remove since the file no longer exists.
 2018-04-03 08:04:25 -05:00
Will Dietz dee77f7e43 git: 2.16.3 -> 2.17.0 2018-04-02 16:40:08 -05:00
Will Dietz 503f8efbcd epoxy: explicitly search libGL path as fallback 
Don't rely on questionable impact of DT_RPATH on dlopen().

This is a bit of a messy subject, but probably the clearest
reference to motivate *not* relying on how dlopen() behaves
in the presence of RPATH or RUNPATH is the following:

https://sourceware.org/ml/libc-hacker/2002-11/msg00011.html

FWIW the dlopen() manpage only mentions the the RPATH
and RUNPATH in the "executable file for the calling program";
no mention of the executable files for libraries--
this has been brought to the attention of the relevant
parties and AFAICT nothing has been done.

The best reference for glibc behavior is
apparently to ... "try it and see".
Luckily a generous soul did exactly that
and reported the findings:

https://www.spinics.net/lists/linux-man/msg02291.html

Qt wrote on the subject a bit when they were bit by this,
linking to the above articles (directly or indirectly).

See:
http://blog.qt.io/blog/2011/10/28/rpath-and-runpath/

--------

Since we know the path of libGL at build-time for libepoxy,
there's a simple solution we can use to avoid all of this:
simply teach libepoxy to explicitly look in the libGL path.

This commit patches libepoxy to accomplish this,
looking to "LIBGL_PATH" as a fallback if it cannot find
the libraries otherwise.

---------

This fixes use of libepoxy w/musl on NixOS!
 2018-04-02 12:35:37 -05:00
Will Dietz ddd10e0836 nixos testing: significantly reduce deps needed for OCR tests 
Instead of imagemagick built with many libraries
(notably librsvg which these days requires rust)
use imagemagick_light with support for libtiff added.
 2018-03-31 00:24:15 -05:00
Will Dietz 25b81a053e swift: loosen platforms to all linux other than known-bad i686 2018-03-31 00:19:46 -05:00
Will Dietz a3df96b381 powertop: patch for musl 2018-03-31 00:11:11 -05:00
Will Dietz 09eaddf4eb sddm: add missing include for gettimeofday, fix w/musl 2018-03-31 00:09:02 -05:00
Will Dietz 462ab6d9de icewm: rename strlcat/strlcpy to avoid conflicts, patch to fix w/musl 2018-03-31 00:06:01 -05:00
Will Dietz 26f537ffab swift: 4.0.3 -> 4.1 2018-03-30 23:37:04 -05:00
Will Dietz 2367250cfb libglvnd: patch from upstream for libdl 2018-03-30 23:27:07 -05:00
Will Dietz 1661c0b2c6 go: 1.10 -> 1.10.1 2018-03-29 17:17:12 -05:00
Will Dietz 166a632124 go: 1.9.4 -> 1.9.5 2018-03-29 17:15:57 -05:00
Will Dietz e022366fb7 busybox: 1.28.1 -> 1.28.2 2018-03-28 23:47:13 +01:00
Will Dietz 5ed0ae686b slim: musl patch 2018-03-27 15:14:23 -05:00
Will Dietz ee9ad96898 parted: add missing fnctl include, fix w/musl 
Disable tests on musl, maybe fix or refine this in the future.
 2018-03-27 15:00:29 -05:00
Will Dietz c787555b94 firmwareLinuxNonfree: 2018-01-04 -> 2018-03-20 2018-03-27 13:08:17 -05:00
Will Dietz e78900213b grub2: extend platforms to include all linux 
This works (and is needed) on musl-based linux too!
 2018-03-26 15:53:38 -05:00
Will Dietz 4037d68db4 virglrenderer: fix missing include 2018-03-26 15:32:00 -05:00
Will Dietz 23ace05628 usbredir: -Wno-error 2018-03-26 15:21:20 -05:00
Will Dietz 06e093c270 xfsprogs: replace glibc-2.27 patch w/upstream variant 
This version also works w/musl :)

https://git.kernel.org/pub/scm/fs/xfs/xfsprogs-dev.git/patch/?id=8041435de7ed028a27ecca64302945ad455c69a6
 2018-03-26 14:53:50 -05:00
Will Dietz 69ca33057e xar, libselinux: use musl-fts on musl, fix 2018-03-26 14:00:11 -05:00
Will Dietz ebf04d83c0 musl-fts: add setup-hook to add -lfts 2018-03-26 13:59:59 -05:00
Will Dietz da0da6177d nilfs-utils: apply upstream patches to fix w/musl 2018-03-26 13:58:44 -05:00
Will Dietz a11afa901f Use npth as "pth" for portability, for now only w/musl. 2018-03-26 13:28:13 -05:00
Will Dietz 5e3e3e985d netcat-openbsd: fix w/musl by providing b64_ntop 
It seems many projects have copies of this code--
and AFAICT all are derived from ISC's BIND[2]
although it does not appear present in recent versions.

This is a curious function as it is not documented
nor part of any standard[1].

[1] https://www.lemoda.net/unix/base64/index.html
[2] for example, glibc's copy:
    https://sourceware.org/git/?p=glibc.git;a=blob;f=resolv/README;h=514e9bb617e710f16126c1474561965a2b35653d;hb=HEAD#l104
 2018-03-26 12:43:50 -05:00
Will Dietz b6b94f0b27 man-db: disable tests w/musl as simplest solution 2018-03-25 21:17:50 -05:00
Will Dietz 9204fa2620 propagate libiconv header from musl too 2018-03-25 21:17:49 -05:00
Will Dietz f623858a82 glib: no one wants --witih-libiconv=gnu 
Previously this was set on FreBSD and SunOS...
 2018-03-25 21:17:49 -05:00
Will Dietz 3bf2002199 celt-0.5.1: disable tests that fail to build due to bad code 
These were fixed upstream, but patches don't apply to
0.5.1 which we need to use for spice.

Instead of maintaining backported patches for very deprecated
software, just don't build the tests in the first place.
 2018-03-25 20:25:01 -05:00
Will Dietz 3675d77058 vde2: musl patch 2018-03-25 19:06:59 -05:00
Will Dietz 969499f3e2 libbsd: more upstream patches 2018-03-25 18:48:28 -05:00
Will Dietz 88ea78745a libbsd: apply upstream patch to fix w/musl 2018-03-25 18:48:26 -05:00
Will Dietz 5faa470da2 btrfs-progs: disable backtrace, fix build w/musl 2018-03-25 18:46:06 -05:00
Will Dietz 55e59e4557 qemu: musl patches 
(cherry picked from commit 1bf8ff49d55fc7dfdd460f3d4f02c148ed2a2b40)
 2018-03-25 18:41:11 -05:00
Will Dietz 990d29282f udisks: extend "we know better than compiler" to more platforms (musl) 2018-03-25 18:09:48 -05:00
Will Dietz 9f7d5981b8 udisks2: 2.1.6 -> 2.1.8 2018-03-25 18:09:43 -05:00
Will Dietz bae5ec0faf zfs: fix hardcoded include path for libtirpc 2018-03-25 18:08:12 -05:00
Will Dietz 47dc9d7b35 tirpc: remove upstreamed patch, fix w/musl 2018-03-25 18:08:05 -05:00
Will Dietz d6ffaac4ac zfs: libtirpc dep on musl 2018-03-25 18:07:59 -05:00
Will Dietz cb30a1b425 wrapper.c: fixup includes to work w/musl 2018-03-25 18:06:02 -05:00
Will Dietz 2e29f46cc4 libpipeline: 1.4.1 -> 1.5.0 2018-03-25 18:03:23 -05:00
Will Dietz 7b44560aa1 musl: provide debug info, separately, don't disable unwind tables 
Also by setting --enable-debug, asm is preprocessed with CFI directives.

Keep unwind tables does not increase size by much,
and helps debuggers and similar make sense of things.
 2018-03-23 12:59:15 -05:00
Will Dietz b59fd70210 gettext: don't add -lintl on musl, either 
gettext won't actually provide libintl unless we go out of
our way to tell it to do so [1][2].

We could add those flags on musl (as I initially did in [3]),
but then we have two different libintl.h files and generally
some confusion about which gettext is being used.

Instead of sorting that out, for now let's just continue on
without gettext providing libintl-- it's worked well enough so far.

Only change that needs to be made, then, is to avoid
adding -lintl on musl since there is no libintl.

[1] https://github.com/pullmoll/void-packages/commit/c739240fd25e8378fe0467cd51ab483427337410
[2] https://gitweb.gentoo.org/repo/gentoo.git/tree/dev-libs/libintl/libintl-0.19.8.1.ebuild?id=332e48712b6521697f992f923c9c985482dd1c36#n41
[3] https://github.com/dtzWill/nixpkgs/commit/729302f29aea8d380196d114e7d2f904add40711
 2018-03-23 03:04:55 -05:00
Will Dietz dcaea58eb7 git: 2.16.2 -> 2.16.3 2018-03-22 18:29:11 -05:00
Will Dietz ebe7b86bac lua-{5.2,5.3}: cross fixups, drop crossAttrs 
mostly just forward environment variables to make arguments,
this partially reverts 5d1e51a199
which removed them because they're already set in env--
but that's not enough to override make vars.

Also, readline is buildInput not nativeBuildInput

(we need headers and to link against it)
 2018-03-21 16:13:27 -05:00
Will Dietz 0fdc4cb94b ngx_brotli: update for fixes, zero copy. 2018-03-21 15:15:50 -05:00
Will Dietz f967d23403 erlang R18-20: updates 2018-03-21 11:12:14 -05:00
Will Dietz a7bc556ce9 gtkspell3: add missing isocodes dep 2018-03-21 10:21:42 -05:00
Will Dietz 63ba3490de pan: enable spell-checking features by default 
Seems to work, not sure why disabled.

Steps are something like:
* install dictionaries into profile
* attempt to compose a message
  * (navigate dialogs about server and posting profile)
* enable spell-checking from edit->checkbox
  (note this toggle is avail even when pan can't spellcheck)
 2018-03-21 10:18:21 -05:00
Will Dietz 7d873b8fd2 pan: ensure features we provide deps for are used, move to gtk3 2018-03-21 09:48:23 -05:00
Will Dietz 96f0d4e98c pan: wrap to provide gpg2, which it crashes without 
Fixes #37556.
 2018-03-21 09:47:55 -05:00
Will Dietz 8265979114 pan: 0.139 -> 0.144 2018-03-21 07:50:40 -05:00
Will Dietz 00fb7d29bf vim-plugins: updates! 
Only updated the ones I have experience with,
so this is not exhaustive.
 2018-03-20 15:48:50 -05:00
Will Dietz 8727284a4b haskell: only use glibcLocales when using glibc 
Fixes eval w/musl.

Possible alternative would be using glibcLocales
when it is non-null, to avoid duplicating the logic.
 2018-03-20 14:38:40 -05:00
Will Dietz 9aa22191cf python*: set thread stack size on musl 
Ensure recursion limit is reached before stack overflow.

Python does this for OSX and BSD:
https://github.com/python/cpython/blob/13ff24582c99dfb439b1af7295b401415e7eb05b/Python/thread_pthread.h#L22

Size of 1MB chosen to match value in Alpine:
https://git.alpinelinux.org/cgit/aports/commit/main/python2/APKBUILD?id=2f35283fec8ec451fe5fb477dd32ffdcc0776e89

Manual testing via Alpine's test-stacksize.py crashes on these
previously, and works with these changes.
 2018-03-20 08:14:04 -05:00
Will Dietz 48c3e60061 aws-sdk-cpp: 1.4.10 -> 1.4.15 2018-03-19 20:05:53 -05:00
Will Dietz d01b1f6b1d numpy: fix on musl 2018-03-15 18:39:59 -05:00
Will Dietz c292469c6f numpy: 1.14.1 -> 1.14.2 2018-03-15 18:39:48 -05:00
Will Dietz 837538da43 openblas: don't build w/openmp on musl 2018-03-15 18:27:22 -05:00
Will Dietz e6ff717b56 speech-tools: run tests 2018-03-15 01:29:06 -05:00
Will Dietz 92bd681de1 speech-tools: is __isnanf, use portable isnan (c99) 2018-03-15 01:23:22 -05:00
Will Dietz ce86490530 speech-tools: 2.4 -> 2.5.0 2018-03-15 01:11:10 -05:00
Will Dietz b2e6952f37 llvm4: same treatment 2018-03-14 20:04:38 -05:00
Will Dietz 60bf299e6a llvm6: patch to enable sanitizers w/musl 2018-03-14 20:04:21 -05:00
Will Dietz a8824ab4fb llvm_5: patch sanitizers to enable w/musl 
https://gitweb.gentoo.org/proj/musl.git/plain/sys-libs/compiler-rt-sanitizers/files/compiler-rt-sanitizers-5.0.0-musl-patches.patch
 2018-03-14 20:04:20 -05:00
Will Dietz 99131b4fc2 pycurl: 7.19.5.1 -> 7.43.0.1 2018-03-14 17:50:25 -05:00
Will Dietz b546c7440e busybox: patchshebangs to fix build on "limited" /bin/sh builders 2018-03-14 16:58:54 -05:00
Will Dietz 4bc0f88bb3 libdwarf: 20170709 -> 20180129 
Upstream describes changes:

Fixes libdwarf/dwarfdump vulnerabilities related to detecting corrupt
DWARF and includes other small improvements
 2018-03-14 15:42:45 -05:00
Will Dietz f911edaf43 llvm6: Build with support for the experimental WASM backend 2018-03-14 15:35:24 -05:00
Will Dietz 77d794fe64 closurecompiler: 20170910 -> 20180204 2018-03-14 15:08:21 -05:00
Will Dietz 43c474457c gllvm: init at 2018-02-09 2018-03-14 14:50:17 -05:00
Will Dietz 2184aac78c wllvm: 1.1.3 -> 1.1.5 2018-03-14 14:29:10 -05:00
Will Dietz 273fd896bc virtualbox: 5.2.6 -> 5.2.8 
Tested against all the VirtualBox VM tests.

Signed-off-by: aszlig <aszlig@nix.build>
Closes: #36127
 2018-03-13 22:07:25 +01:00
Will Dietz 29d46471e0 metrics: fix "values" to be count not bytes, to match prev behavior 
https://github.com/NixOS/nixpkgs/pull/36896#discussion_r174163744
 2018-03-13 13:29:15 -05:00
Will Dietz 5eb07e2396 metrics: fixup sed expr to work with Nix 2 2018-03-12 22:49:17 -05:00
Will Dietz 708e3fb160 zed: inherit libtool, fix eval 2018-03-12 22:47:47 -05:00
Will Dietz c2c9d393a8 runc: 1.0.0rc4 -> 1.0.0rc5 
https://github.com/opencontainers/runc/releases/tag/v1.0.0-rc5
 2018-03-11 15:42:36 -05:00
Will Dietz 54c4c183dd glibc: make fetchpatch optional ("? null"), just in case. 2018-03-11 11:16:53 -05:00
Will Dietz 82644c15e5 glibc: patch to fix building w/musl-based stdenv 
See:
https://sourceware.org/bugzilla/show_bug.cgi?id=21604
 2018-03-11 10:53:31 -05:00
Will Dietz 3d88d7af0f bash: 4.4p12 -> 4.4p19 2018-03-09 13:11:07 -06:00
Will Dietz 42c33ce12f llvmPackages_6: build using gcc6 on i686, like others. 2018-03-08 13:03:22 -06:00
Will Dietz b36ea92f88 llvm6: rc3 -> final 2018-03-08 10:01:35 -06:00
Will Dietz 08db02e37a llvm6: rc2 -> rc3 2018-03-07 19:03:10 -06:00
Will Dietz 9306e2fb6f llvm6: remove copied "remove this on next rebuild" value :) 2018-03-07 19:03:09 -06:00
Will Dietz c3e16a42ee llvm6: drop perl and groff, as in llvm4/5 2018-03-07 19:03:09 -06:00
Will Dietz e876e97ef0 llvm6: musl compat 2018-03-07 19:03:08 -06:00
Will Dietz ab0a72a33d clang-6 test variants 2018-03-07 19:03:08 -06:00
Will Dietz 62011cfaea llvm6: rc1 -> rc2 2018-03-07 19:03:07 -06:00
Will Dietz 11cb7d00ca llvm6: same python output fix from #33871 
Unlike that PR, however, this doesn't need to go to staging
since no one depends on a package that doesn't exist yet :).
 2018-03-07 19:03:07 -06:00
Will Dietz 0d90539cda llvm6: no need to patch on aarch64, already applied 2018-03-07 19:03:06 -06:00
Will Dietz a192c9d67a lld: Add dep on libxml2 2018-03-07 19:03:06 -06:00
Will Dietz 42c3a74a2a llvm6: rc1, hashes, top-level 2018-03-07 19:03:05 -06:00
Will Dietz 4463de6ec7 llvm6: copy llvm5 2018-03-07 19:03:05 -06:00