partial fix for Issue 5570 #3372
Conversation
|
Hmm, does EXTRA_C_SOURCES work? |
|
Isn't argtypes a visitor yet? I was looking forward to booting the entire file from gdc dfrontend sources. |
|
@WalterBright There is no @ibuclaw Yes it is. |
| case 15: | ||
| if (!global.params.is64bit) | ||
| goto Lmemory; | ||
| t1 = NULL; |
There was a problem hiding this comment.
t1 is already null by the looks of it.
There was a problem hiding this comment.
Yes, but I put that in because the other cases set t1.
|
Why the fail on win64? |
|
I disabled the Win64 tests for now. The Win64 stuff is totally different, and these tests had never been run on it before, so this is not a regression. |
|
Does that mean that the win64 abi code gen is also broken and this bug report is incorrect in marking it as linux only? And that this fix is even more partial? I'm just asking so that we don't loose track of additional bugs that might have just been revealed. And any test that's disabled due to bugs should have a bug filed (probably with a severity in the blocker, critical, or major range) and a comment in the test code pointing at that bug report, and probably should emit a "disabled on xx due to yy" sort of comment. Anything to draw attention to a disabled test. Otherwise, it'll just get lost in the noise of all the open bugs. |
|
@braddr yes it means that with some structs the Win64 ABI is not followed. The Win64 ABI is very, very different and would need a completely different fix. I think it merits a separate bugzilla issue. http://d.puremagic.com/issues/show_bug.cgi?id=12343 |
|
Thanks! This is helping FB Thrift work. |
|
This pull request introduced a regression https://issues.dlang.org/show_bug.cgi?id=15455 |
macOS: add core.sys.darwin.fcntl
Partial in that it doesn't solve the problem for returning structs of sizes 3,5-7,9-15
https://d.puremagic.com/issues/show_bug.cgi?id=5570