| 1 |
Rich Freeman <rich0@g.o> wrote: |
| 2 |
> |
| 3 |
> I tend to agree. I've been using stable gcc with -flto in my CFLAGS |
| 4 |
> for a while now with only isolated problems. |
| 5 |
|
| 6 |
I wouldn't call these problems isolated: |
| 7 |
My current exception file has 340 lines, some of them containing |
| 8 |
wildcards, and it has a tendency to grow. |
| 9 |
For comparison: I have ~1400 packages installed. |
| 10 |
Maybe our milage varies because I have many multimedia packets |
| 11 |
which use lots of libraries - these usually break. |
| 12 |
|
| 13 |
> I wouldn't be surprised if some of them now work. |
| 14 |
|
| 15 |
Not much change between gcc versions up to 4.8: |
| 16 |
Some packages worked with newer versions some others broke instead. |
| 17 |
I have no experience yet with 4.9. |
| 18 |
|
| 19 |
> Anybody who is using it should be prepared to run into the odd |
| 20 |
> breakage. |
| 21 |
|
| 22 |
That's why it is wise that gentoo does not recommend to use LTO |
| 23 |
on a global scale. |
| 24 |
However, for packages which are tested by upstream with LTO...? |
| 25 |
|
| 26 |
> It does make sense to filter the flag when it is known to |
| 27 |
> not work. |
| 28 |
|
| 29 |
This would be the best solution of course: Recommend LTO and |
| 30 |
filter every occassion which breaks. But currently this is |
| 31 |
not realistic, because too many ebuilds would need to be tested |
| 32 |
and checked. Moreover, sometimes it depends on the gcc version |
| 33 |
whether filtering is necessary (although, as mentioned, these |
| 34 |
cases are relatively rare with <gcc-4.9). |
| 35 |
So, one should not expect this in any near future. |
Archives