Switch linkage of ___tls_get_addr to extern(C).#2034
Switch linkage of ___tls_get_addr to extern(C).#2034dlang-bot merged 1 commit intodlang:masterfrom ibuclaw:externctls
Conversation
|
Thanks for your pull request, @ibuclaw! Bugzilla referencesYour PR doesn't reference any Bugzilla issue. If your PR contains non-trivial changes, please reference a Bugzilla issue or create a manual changelog. Testing this PR locallyIf you don't have a local development environment setup, you can use Digger to test this PR: dub fetch digger
dub run digger -- build "master + druntime#2034" |
|
Rebased now that dlang/dmd#7620 is in, lets see what the test-suite has to say... |
|
I guess today isn't your lucky day - Darwin_64_32 fails with: |
|
Does someone have a box I can use for debugging? |
|
Not for osx? |
|
Whoops, I got carried away. Yes, please rebase. |
|
I imagine that we still need to work out what is wrong with the compiler first. ;-) |
|
I have an OSX box and could narrow the problem, or we could just ditch 32bit OSX :) |
Apple deprecate support for 32-bit a long time ago and they will remove support for it completely with the next version:
https://support.apple.com/en-us/HT208436 So please let's not waste time on a dead horse. |
Yes please. But there might be necessary to convenience some other people. |
|
OSX32-bit support was dropped can this be closed or merged? |
I would guess so. Anyone else? |
|
There's still the compiler component to remove. |
|
Does that need to be done before this? |
|
I wouldn't have thought so. From memory, the compiler side is just a special case for recognizing and not mangling symbols named |
So if it's |
|
I'm rebasing to have another look at what the error is. Switching from extern(D) to extern(C) would change the ABI on 32bit (pointer passed in register vs. pushed on stack), maybe there's something else in the compiler that needs to change in lock-step. |
|
Great, the testsuite passes now because 32bit Darwin is disabled... |
|
Might as well merge this then? I'll see if I can track down possible OSX 32-bit problems in the teardown PR that will be sent to dmd. What I'll be notably looking for it anything that suggests the D calling convention around ___tls_get_addr. |
Rename to __tls_get_addr to keep same ABI.
This depends on dlang/dmd#7620 getting merged.