Commit Graph

511 Commits

Author SHA1 Message Date
Kamil Dudka
59a39fa268 apply patches automatically to ease maintenance 2016-12-02 13:48:08 +01:00
Kamil Dudka
8d346246a7 disable the test-lock gnulib test prone to deadlock
See https://lists.gnu.org/archive/html/bug-gnulib/2015-07/msg00033.html
and https://libvirt.org/git/?p=libvirt.git;a=commitdiff;h=25fb4c6 for
some context.
2016-12-02 11:08:45 +01:00
Kamil Dudka
8d9eac4093 new upstream release 8.26 2016-12-01 19:14:56 +01:00
Kamil Dudka
3ed22ec3e8 md5sum,sha*sum: fix --ignore-missing with checksums starting with 00 2016-10-31 17:52:01 +01:00
Tomas Mraz
21de8b5251 rebuild with OpenSSL 1.1.0 2016-10-11 13:25:50 +02:00
Kamil Dudka
380647993e Resolves: #1365933 - ls: allow interruption when reading slow directories 2016-09-07 20:17:56 +02:00
Kamil Dudka
0e3a55d8dc coreutils.spec: do not compress files that are not installed 2016-09-07 20:09:15 +02:00
Kamil Dudka
39f7c7c836 run autoreconf in %prep 2016-07-19 15:13:06 +02:00
Kamil Dudka
adf19d29ec drop post-install fix for Japanese locales that no longer applies 2016-07-19 14:00:38 +02:00
Kamil Dudka
8b01f2371c Resolves: #1355780 - fix 'sort -h -k' in locales that use blank as thousands separator 2016-07-19 13:53:33 +02:00
Kamil Dudka
6cf6cd48dd make 'sort -h' work for arbitrary column
... even when using UTF-8 locales
2016-07-14 16:50:51 +02:00
Kamil Dudka
9551d101b0 Related: #1335320 - do not print DIR_COLOR to stdout 2016-07-11 22:00:42 +02:00
Kamil Dudka
fe09c9cd04 Related: #1354078 - drop %pre dependency on install-info 2016-07-11 17:36:22 +02:00
Kamil Dudka
3bf1e52ff3 Resolves: #1339135 - install -Z now sets default SELinux context for created directories 2016-07-11 17:27:02 +02:00
Kamil Dudka
3792f36ae5 Resolves: #1354078 - drop the %pre scriptlet
... which is no longer needed.  Fileutils, Shellutils, and Textutils
were merged into GNU Coreutils in 2002.  It should be safe not to handle
the transition now in 2016.
2016-07-11 17:17:36 +02:00
Kamil Dudka
e207ff79a3 Resolves: #1349579 - clarify recognition of "^COLOR.*none" in /etc/DIR_COLORS 2016-07-11 17:14:09 +02:00
Kamil Dudka
7d9e8632f3 coreutils.spec: fix the last change log entry 2016-07-11 17:13:20 +02:00
Jakub Martisko
0a63fa44ae (un)expand: UTF8-BOM header detection 2016-07-07 12:53:26 +02:00
Ondřej Vašík
34ffa2ef55 change way of detection of interactive shell in colorls.sh script 2016-06-24 12:45:22 +02:00
Kamil Dudka
ee3bfc24ca add BR for glibc-langpack-en to prevent the expand/mb test from failing 2016-06-20 16:57:49 +02:00
Kamil Dudka
0433f98b1b coreutils.spec: sort BuildRequires alphabetically
No observable changes intended.  Just to ease the maintenance.
2016-06-20 16:49:30 +02:00
Kamil Dudka
3b6df16a3d Resolves: #1348043 - do not use /bin/mv in %post
... to avoid a circular dependency.

The code where /bin/mv was used does not seem to be useful any more.
Fileutils, Shellutils, and Textutils were merged into GNU Coreutils
in 2002.  It should be safe not to handle the transition now in 2016.
2016-06-20 16:40:30 +02:00
Kamil Dudka
46ebf91d5f Related: #1335320 - maintain downstream DIR_COLOR* files as a patch
... against src/dircolors.hin rather than full copies of the file.
There should be no user-visible change introduced by this commit.

This is just to ease the maintenance -- the resulting files cannot
diverge from upstream without any notice.
2016-06-17 17:14:24 +02:00
Kamil Dudka
50e40c5f2b Resolve: #1335320 - sync /etc/DIR_COLORS with latest upstream 2016-06-17 17:13:50 +02:00
Kamil Dudka
28edec2fbc make sure that the license file is installed
... even if coreutils-common is not
2016-06-15 17:32:29 +02:00
Kamil Dudka
8236de4f4b handle info doc in RPM scriptlets of coreutils-common
... which provides it
2016-06-15 17:31:56 +02:00
Jakub Martisko
ffecc6b864 (un)expand: fix regression in handling of input files, where only the first file was processed. 2016-06-14 12:47:41 +02:00
Pádraig Brady
45ce637bf4 maint: remove build workarounds for old upstream bugs
* coreutils.spec: coreutils 8.25 fixed a couple of issues,
so remove our workarounds.
2016-04-22 14:14:54 +01:00
Ondřej Vašík
b0302efc50 adjust old i18n patch cut support for coreutils-8.25 2016-03-05 15:53:55 +01:00
Ondřej Vašík
96a98c054d Disable new i18n cut support 2016-03-05 15:28:11 +01:00
Ondřej Vašík
06a5ca2716 cut: fix regression in handling fields for lines wider than 64 chars (#1304839) 2016-02-15 12:47:08 +01:00
Lubomir Rintel
593569e76b Fix the unexpand 2016-02-11 17:56:10 +01:00
Ondřej Vašík
5440a2d9db Add new base32 binary to the list of utils 2016-02-02 15:08:35 +01:00
Ondrej Oprala
ab00e1ea64 Adjust the i18n patch for coreutils-8.25 2016-01-22 14:58:03 +01:00
Ondřej Vašík
7d9c9afa38 Initial commit for coreutils 8.25 ... still need to fix two failing tests before build (likely i18n stuff) 2016-01-20 16:43:39 +01:00
Ondrej Oprala
11e5aa1d55 cut: be MB for ALL archs 2016-01-15 11:02:13 +01:00
Ondrej Oprala
01067b2813 Use the new i18n implementation for the cut utility 2016-01-15 10:20:05 +01:00
Ondřej Vašík
452aa4a7e3 mv: prevent dataloss when source dir is specified multiple times (#1297464, by P.Brady) 2016-01-13 12:30:46 +01:00
Pádraig Brady
c553cab787 give explicit priority to coreutils over coreutils-single
Make the main coreutils package Obsoletes: coreutils-single
so that it has priority in depsolving.
2015-12-14 14:14:33 +00:00
Pádraig Brady
63c57dca19 reduce dependencies for coreutils-single
Don't depend on libcrypto from the multicall binary in
the coreutils-single package, because that trades space for speed.

Don't depend on libgmp from the multicall binary in
the coreutils-single package, because that only benefits
factor and expr.  The large size of gmp is not seen as
an appropriate tradeoff for this functionality.

Note also there is reduced startup overhead for all tools
with these removed due to not linking with the shared libs.
2015-12-04 11:04:24 +01:00
Pádraig Brady
7f1720d9a2 fix inclusion of /usr/bin/kill in coreutils-single
we need to remove /usr/bin/kill.single
2015-12-03 19:40:06 +00:00
Pádraig Brady
1f6f388ba3 avoid warning about twice specified %files 2015-12-03 19:38:51 +00:00
Ondrej Oprala
32b1e5a154 Use the new i18n implementation for expand/unexpand 2015-12-01 11:20:49 +01:00
Ondřej Vašík
8e9b3fd0aa Merge conflicts 2015-11-30 14:33:26 +01:00
Ondřej Vašík
49c29b4069 Resolves:#1286338 - coreutils-single should provide versioned coreutils 2015-11-30 14:32:27 +01:00
Pádraig Brady
5dc61e9dc1 maint: update stale comments
The comments re %post processing were for a non commited change
that created symlinks during %post that clobbered any existing binaries.
Now all files are directly managed by rpm.
2015-11-24 01:44:48 +00:00
Pádraig Brady
5fb9bc4700 split package and provide coreutils-single
Notes about coreutils:

 - Drops from 14.2MB to 5.5MB.
 - Still requires coreutils-common, though being separated
   allows for easier removal if required.
 - Explicitly conflicts with coreutils-single to avoid errant installs

Notes about coreutils-single:

 - Contains a single multicall binary which is 1.2MB
 - Have to force install over any existing coreutils package
    so best used for initial install without existing coreutils package
 - Doesn't require (but suggests) coreutils-common
    so install that manually for docs, translations, and colors etc.

Notes about coreutils-common:

 - 8.7MB
 - Having this split out gives an easy way to remove all
   docs and translations.

Note RemovePathPostfixes: introduced with rpm 4.13
is used to facilitate producing the two sets of conflicting binaries
from the same build, while maintaining appropriate --fileprovide
in each generated rpm.
2015-11-20 08:56:09 +01:00
Pádraig Brady
f0951fecac clean stale parts of the spec file 2015-11-20 08:56:09 +01:00
Kamil Dudka
c7c3ee3fab Resolves: #1259942 - fix memory leak in sort/I18N
Patches written by Pádraig.

Note that the corresponding i18n/sort-month test was not included
because it breaks unless sort is compiled -Dlint and we do not want
to decrease performance of the resulting RPMs (and valgrind is not
installed in production buildroots anyway).
2015-09-16 20:09:55 +02:00
Ondřej Vašík
2ad92d25c0 fix one still existing occurance of non-full path in colorls.sh 2015-09-12 08:00:58 +02:00