-
-
Notifications
You must be signed in to change notification settings - Fork 606
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
fix 32 bit OSX build #5788
fix 32 bit OSX build #5788
Conversation
Great, this is currently blocking the nightly builds. Please |
This bug isn't in stable, it was just recently 2016-05-06 introduced in master. |
And |
@@ -474,7 +474,7 @@ void DtBuilder::cat(DtBuilder *dtb) | |||
/************************************** | |||
* Repeat a list of dt_t's count times. | |||
*/ | |||
void DtBuilder::repeat(dt_t *dt, size_t count) | |||
void DtBuilder::repeat(dt_t *dt, unsigned count) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is there a reason why d_size_t
is not used here as well?
The D declaration of |
@jacob-carlborg I had messed up my files due to a merge, and the wrong files got in this PR. Corrected. |
LGTM 👌 |
This is kind of a workaround. Any plans to properly support linking to a C++ function with a |
I don't think we plan to introduce size_t as distinct ("real") type, so it cannot have a different mangling than uint/ulong. Not sure what else we could do, the information that sth. was a size_t is lost pretty early. |
Nice, working nightlies again http://nightlies.dlang.org/dmd-2016-05-17/. |
If we want to provide the frontend as a library people can use to write tool, we will ultimately have to go for a less / non-destructive approach, and store the full information. |
Why are we building a 32-bit OSX binary at all? |
@yebblies I know Guillaume Piolat, @p0nce, is still using 32bit to build plugins for audio applications on OS X. What I understand is that users of these audio applications have loads of, I assume old, plugins compiled for 32bit. If they are old and not maintained they might not be able to get a 64bit version of the plugins and therefore cannot move to a 64bit version of the audio application. The iOS simulator doesn't emulate ARM, instead it runs x86 binaries. Even though there are iOS devices which are 64bit, the Apple Watch is still 32bit, as far as I understand. Although one might argue that one would use LDC to target iOS devices including the simulator. |
I'm not suggesting dropping OSX32 as a target, just as a host. |
We already have officially dropped OSX32 as a host. But I find it convenient to build it as a way of finding portability issues, and having the compiler be able to build itself is part of testing the OSX32 support. |
Aha, you mean building the compiler as a 32bit binary. No, there's really no point in doing that when you can cross-compile. |
Good question, it's due to the quirkiness of our build script, which was originally written to allow as many configuration options as possible, leading to lots of complicated edge cases. |
Most of the issues it finds seem to be exactly the issues we dropped it for. My motivation was avoiding the ugly hacks and special cases that we're now adding. If we're going to actively maintain it anyway then we might as well have it as an official host platform.
It doesn't have to be, and it isn't on the autotester. It's a weird kind of limbo in between fully supported and unsupported at the moment - but I guess so are solaris and win64. |
size_t has mangling issues on 32 bit OSX