| 1 |
On Sat, Apr 26, 2014 at 11:00 AM, Martin Vaeth <martin@×××××.de> wrote: |
| 2 |
> Rich Freeman <rich0@g.o> wrote: |
| 3 |
>> It does make sense to filter the flag when it is known to |
| 4 |
>> not work. |
| 5 |
> |
| 6 |
> This would be the best solution of course: Recommend LTO and |
| 7 |
> filter every occassion which breaks. But currently this is |
| 8 |
> not realistic, because too many ebuilds would need to be tested |
| 9 |
> and checked. Moreover, sometimes it depends on the gcc version |
| 10 |
> whether filtering is necessary (although, as mentioned, these |
| 11 |
> cases are relatively rare with <gcc-4.9). |
| 12 |
> So, one should not expect this in any near future. |
| 13 |
|
| 14 |
Sounds like many are already using it, and thus much of this testing |
| 15 |
has already been done. The results simply need to be collected. |
| 16 |
|
| 17 |
FWIW the list of packages I have issues with include: |
| 18 |
app-antivirus/clamav |
| 19 |
app-backup/bacula |
| 20 |
app-office/libreoffice |
| 21 |
dev-libs/elfutils |
| 22 |
dev-libs/libaio |
| 23 |
dev-libs/libpcre |
| 24 |
dev-python/pygtkglext |
| 25 |
dev-python/wxpython |
| 26 |
dev-qt/qt-creator |
| 27 |
dev-tex/luatex |
| 28 |
dev-util/kdevelop |
| 29 |
dev-util/kdevplatform |
| 30 |
kde-base/gwenview |
| 31 |
kde-base/kdelibs |
| 32 |
kde-base/korganizer |
| 33 |
media-gfx/inkscape |
| 34 |
media-sound/amarok |
| 35 |
media-video/mplayer |
| 36 |
net-libs/webkit-gtk |
| 37 |
net-misc/nx |
| 38 |
net-wireless/gnuradio |
| 39 |
sys-apps/kmod |
| 40 |
sys-devel/llvm |
| 41 |
sys-power/upower |
| 42 |
x11-libs/gtkglext |
| 43 |
x11-libs/wxGTK |
| 44 |
|
| 45 |
There are also other packages which I build with very simplified |
| 46 |
CFLAGS (just -O2 basically) which may or may not have LTO issues. |
| 47 |
Also, some packages above like libreoffice may very well build just |
| 48 |
fine if you have sufficient RAM. |
| 49 |
|
| 50 |
I do run a modest number of multimedia packages, and only the ones |
| 51 |
above have caused issue (though if any haven't been updated in a few |
| 52 |
years by some stroke of luck and being abandoned then they wouldn't be |
| 53 |
tested with LTO). |
| 54 |
|
| 55 |
Rich |
Archives