| 1 |
> On 8 Nov 2021, at 11:18, Michał Górny <mgorny@g.o> wrote: |
| 2 |
> Hi, |
| 3 |
> A few years back I've slotted LLVM and Clang to make the life with |
| 4 |
> revdeps easier. Long story short, every major LLVM release (which |
| 5 |
> happens twice a year) breaks API and it takes some time for revdeps to |
| 6 |
> adjust. Slotting made it possible to install multiple versions |
| 7 |
> simultaneously, and therefore let "faster" packages use newer LLVM |
| 8 |
> without being blocked by "slower" packages on the user's system. |
| 9 |
> |
| 10 |
> Unfortunately, this ended up pretty bothersome to maintain. Besides |
| 11 |
> making ebuilds quite complex (and prone to mistakes), I'm hearing more |
| 12 |
> and more reports of programs being broken through getting multiple LLVM |
| 13 |
> versions in the link chain. |
| 14 |
|
| 15 |
I think this might just be Blender and friends which are especially fragile. |
| 16 |
|
| 17 |
We may be able to get away with just coordinating those together. |
| 18 |
|
| 19 |
> WDYT? |
| 20 |
|
| 21 |
If we can help it, I'd really really prefer we don't. Being able to test |
| 22 |
various different various of Clang quickly (just like gcc) is really helpful. |
| 23 |
|
| 24 |
(especially given one day, we might dare to dream of using Clang |
| 25 |
for the system toolchain. It becomes a lot easier to check for |
| 26 |
regressions if you can just flip the version.) |
| 27 |
|
| 28 |
Best, |
| 29 |
sam |
Archives