Log of /eclass/flag-o-matic.eclass
Parent Directory
| 
Revision Log
Revision
1.198 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Sat Jun 14 05:14:55 2014 UTC
(4 years ago)
by
robbat2
Branch:
MAIN
Changes since
1.197: +5 -1 lines
Diff to
previous 1.197
,
to
selected 1.66
If you an empty argument to append-libs, you end up with a dangling -l without a library. Catch this in QA. Caught infra with a broken ncurses in net-dialup/xc with append-libs "$($(tc-getPKG_CONFIG) --libs ncurses)"
Revision
1.194 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Fri Dec 27 21:39:36 2013 UTC
(4 years, 5 months ago)
by
robbat2
Branch:
MAIN
Changes since
1.193: +15 -9 lines
Diff to
previous 1.193
,
to
selected 1.66
Per discussion with Flameeyes, make -l and -L always valid, and only warn about other arguments to append-libs. Also document expected arguments to append-libs.
Revision
1.193 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Fri Dec 27 21:27:38 2013 UTC
(4 years, 5 months ago)
by
robbat2
Branch:
MAIN
Changes since
1.192: +7 -2 lines
Diff to
previous 1.192
,
to
selected 1.66
There are usages in the tree of "append-libs $(pkg-config ...)"; if pkg-config returns something other than a library, append-libs can end up inserting -l-L/usr/lib64 or other weirdness. We need to check for these and warn about them.
Revision
1.189 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Mon Sep 30 06:39:56 2013 UTC
(4 years, 8 months ago)
by
vapier
Branch:
MAIN
Changes since
1.188: +24 -9 lines
Diff to
previous 1.188
,
to
selected 1.66
append-*flags: revert to old behavior where we always append the flags specified rather than doing testing on them; it is not feasible to handle flags that have spaces in them #417047
Revision
1.184 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Thu Nov 29 23:27:32 2012 UTC
(5 years, 6 months ago)
by
vapier
Branch:
MAIN
Changes since
1.183: +2 -2 lines
Diff to
previous 1.183
,
to
selected 1.66
test-flag-PROG: read from stdin rather than directly from /dev/null as some compiler driver versions cannot handle that #443778
Revision
1.182 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Sun Nov 18 08:27:00 2012 UTC
(5 years, 7 months ago)
by
vapier
Branch:
MAIN
Changes since
1.181: +8 -7 lines
Diff to
previous 1.181
,
to
selected 1.66
test-flags: set the lang according to the compiler we are using rather than always testing the C frontend so that we can handle language-specific flags correctly #439914 by Alphat-PC
Revision
1.168 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Mon Jan 16 20:03:32 2012 UTC
(6 years, 5 months ago)
by
vapier
Branch:
MAIN
Changes since
1.167: +4 -4 lines
Diff to
previous 1.167
,
to
selected 1.66
quote single flag arg to test-flag-PROG since we are only testing one flag, and we want to handle any weirdness in that flag (such as spaces) #399055
Revision
1.162 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Wed Dec 28 05:48:28 2011 UTC
(6 years, 5 months ago)
by
dirtyepic
Branch:
MAIN
Changes since
1.161: +2 -10 lines
Diff to
previous 1.161
,
to
selected 1.66
Remove UNSTABLE_FLAGS. Keyword status should not determine what flags are used. Move -Os to allowed flags as it's been stable since 4.0ish and drop -O0 because it breaks things a lot.
Revision
1.157 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Fri Nov 11 07:06:25 2011 UTC
(6 years, 7 months ago)
by
dirtyepic
Branch:
MAIN
Changes since
1.156: +6 -6 lines
Diff to
previous 1.156
,
to
selected 1.66
Test that appended flags are valid. This allows people to add flags that were unsupported in earlier releases without needing to do version checking.
Revision
1.152 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Tue Jul 12 04:08:52 2011 UTC
(6 years, 11 months ago)
by
vapier
Branch:
MAIN
Changes since
1.151: +3 -3 lines
Diff to
previous 1.151
,
to
selected 1.66
use -c rather than -S when testing flags so that we can check for assembler flags and not just compiler flags (which lets use check -Wa,--xxx flags #373743)
Revision
1.124 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Thu Jul 3 05:30:54 2008 UTC
(9 years, 11 months ago)
by
dberkholz
Branch:
MAIN
Changes since
1.123: +83 -12 lines
Diff to
previous 1.123
,
to
selected 1.66
[Fortran] Add Fortran support to filter-flags(), append-flags(), is-flagq(), strip-flags(), strip-unsupported-flags(), get-flag(). Add new functions parallel to existing ones for Fortran: append-fflags(), test-flag-F77(), test-flags-FC().
Revision
1.123 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Sat Jun 7 16:42:32 2008 UTC
(10 years ago)
by
flameeyes
Branch:
MAIN
Changes since
1.122: +7 -1 lines
Diff to
previous 1.122
,
to
selected 1.66
Warn when using append-ldflags to pass libraries to link to.
Libraries shouldn't be passed as LDFLAGS, the flags should be passed
before the object files to link, while libraries should be passed
after. Passing libraries through LDFLAGS is one of the causes of
--as-needed failures.
Revision
1.98 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Fri Dec 9 18:40:11 2005 UTC
(12 years, 6 months ago)
by
fmccor
Branch:
MAIN
Changes since
1.97: +2 -2 lines
Diff to
previous 1.97
,
to
selected 1.66
In replace-sparc64-cflags, we need local 'SPARC64_CPUS="ultrasparc3 ultrasparc v9"'.
Without the ultrasparc3 at the beginning of the list, in some cases, we can end up
with '-mcpu=ultrasparc3' ==> '-mcpu=v83', and subsequently, package configure can
fail. Problem arose from mozilla and a particular set lf CFLAGS.
Revision
1.95 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Thu Oct 13 05:33:34 2005 UTC
(12 years, 8 months ago)
by
vapier
Branch:
MAIN
Changes since
1.94: +10 -10 lines
Diff to
previous 1.94
,
to
selected 1.66
need to declare x as a local var in strip-unsupported-flags so we dont pollute callers env ... also touchup syntax there and in has_m32
Revision
1.92 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Mon Aug 1 10:50:55 2005 UTC
(12 years, 10 months ago)
by
kevquinn
Branch:
MAIN
Changes since
1.91: +26 -2 lines
Diff to
previous 1.91
,
to
selected 1.66
deprecate has_hardened/pic/pie/ssp/ssp_all in favour of gcc-specs-*
from toolchain-funcs
modify has_ssp/has_ssp_all to also try gcc-spec-ssp (for bug #90311
and bug #100902)
Revision
1.66 -
(
view)
(
download)
(
annotate)
-
[selected]
Wed Jul 28 02:18:38 2004 UTC
(13 years, 10 months ago)
by
lv
Branch:
MAIN
Changes since
1.65: +4 -1 lines
Diff to
previous 1.65
cockblock amd64 from the -fno-pic logic, as it will randomly break stuff
Revision
1.64 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Thu Jul 22 15:29:10 2004 UTC
(13 years, 11 months ago)
by
agriffis
Branch:
MAIN
Changes since
1.63: +29 -14 lines
Diff to
previous 1.63
,
to
selected 1.66
update replace-flags to use atom code from filter-flags, related to #42006. Also add ${IUSE} to IUSE for the sake of portage .50 users, which is basically all of them at the moment
Revision
1.58 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Thu Jun 10 15:28:56 2004 UTC
(14 years ago)
by
agriffis
Branch:
MAIN
Changes since
1.57: +76 -78 lines
Diff to
previous 1.57
,
to
selected 1.66
- Localize variables in functions
- Rewrite test_flag to honor ${CC}
- Add function test_version_info to run ${CC} --version and look for
patterns
- Use printf instead of echo in situations where flags are being
printed. This avoids the possibility that echo will interpret the
flags, for example -n
- Modify has_hardened, has_pic, has_pie to use test_version_info.
This also fixes mistakes in ${var/pattern} expressions; many
patterns in this eclass were erroneously written to be regular
expressions instead of globs.
- Modify has_m64, has_m32 to use test_flag
Revision
1.56 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Thu Jun 10 00:24:57 2004 UTC
(14 years ago)
by
lv
Branch:
MAIN
Changes since
1.55: +22 -1 lines
Diff to
previous 1.55
,
to
selected 1.66
added has_m64 and has_m32 to flag-o-matic. this should make checking for multilib as simple as has_m64 && has_m32 on all multilib archs other than mips.
Revision
1.51 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Wed May 26 18:53:44 2004 UTC
(14 years, 1 month ago)
by
lv
Branch:
MAIN
Changes since
1.50: +3 -2 lines
Diff to
previous 1.50
,
to
selected 1.66
add -m32 to the list of allowed flags on x86 so that -march=k8 and -mtune=k8 can be used on amd64 machines in 32bit without accidentally producing 64bit code
Revision
1.37 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Sat Feb 21 07:19:29 2004 UTC
(14 years, 4 months ago)
by
vapier
Branch:
MAIN
Changes since
1.36: +19 -5 lines
Diff to
previous 1.36
,
to
selected 1.66
add replace-cpu-flags to help tone down march/mcpu flags
make sure filter-flags/replace-flags catch multiple copies of a flag (for example, if user has -msse -msse -msse and we try to `filter-flags -msse`, we would catch only the first one before)
Revision
1.36 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Sat Feb 21 07:07:56 2004 UTC
(14 years, 4 months ago)
by
vapier
Branch:
MAIN
Changes since
1.35: +73 -52 lines
Diff to
previous 1.35
,
to
selected 1.66
move ALLOWED_FLAGS/UNSTABLE_FLAGS out of global scope since only one or two functions use them
change filter-mfpmath to use other flag-o-matic functions instead of its own code
change instances of $@ to "$@" to provide better quoting
update replace-flags to only work on flag atoms
add more comments for people to understand the code
remove trailing space in etexec-flags/fstack-flags
Revision
1.6 -
(
view)
(
download)
(
annotate)
-
[select for diffs]
Fri Jul 12 15:24:36 2002 UTC
(15 years, 11 months ago)
by
danarmak
Branch:
MAIN
Changes since
1.5: +1 -1 lines
Diff to
previous 1.5
,
to
selected 1.66
add INHERITED variable to all eclasses. this keeps track of what eclasses were inherited. e.g. INHERITED='kde-base kde base'. this will allow drobbins to take eclass mtime changes into account when deciding whether to refresh ebuild cache.
This form allows you to request diffs between any two revisions of this file.
For each of the two "sides" of the diff,
select a symbolic revision name using the selection box, or choose
'Use Text Field' and enter a numeric revision.
| Sticky Tag: |
|
