https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #12 from github-bugzi...@puremagic.com ---
Commits pushed to master at https://github.com/dlang/dmd
https://github.com/dlang/dmd/commit/20ffec71c10a6f6e1651503dbb63cb1a7015cb20
fix issue 18928 - extern(C++) bad codegen on win64, wrong
https://issues.dlang.org/show_bug.cgi?id=18928
Rainer Schuetze changed:
What|Removed |Added
CC||r.sagita...@gmx.de
--- Comment #11 from
https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #10 from Manu ---
Awesome sauce... sadly, for the immediate moment, I really need a fix in DMD
>_<
--
https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #9 from ki...@gmx.net ---
Well I gave it a shot and came up with an ugly hack which might just work, the
first test results are promising:
https://github.com/ldc-developers/ldc/pull/2720
--
https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #8 from Manu ---
So, probability of a prompt fix? Is this just a matter of not having correct
information up front, or is the fix complicated by some issue?
We have an evaluation here that's kinda been shattered by this, but a fast fix
https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #7 from ki...@gmx.net ---
Well it doesn't make too much sense, and even less so to do it for methods
only; I rather think this is some legacy COM compatibility stuff. I did a few
tests, and it really seems to be specific to methods
https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #6 from Manu ---
I recall a few years ago reading some article about MSVC breaking their ABI to
guarantee RVO, because C++11 and move semantics and all that.
I guess it needed to be efficient and therefore justified guaranteeing RVO
https://issues.dlang.org/show_bug.cgi?id=18928
ki...@gmx.net changed:
What|Removed |Added
CC||ki...@gmx.net
--- Comment #5 from
https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #4 from Manu ---
If you look at the disassembly of userdatabase.d, you'll see that the virtual
function is not expecting the string in R8 where VC is putting it. It also
doesn't appear to return via the RVO pointer(?) in RDX.
The static
https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #3 from Manu ---
You'll have to relate these 2 lines to the .cpp file (attached) in main()
Calling code in main():
// static call...
auto x = addUser(test, "test");
7FF74843C587 lea rdx,[string "test" (07FF748448D44h)]
https://issues.dlang.org/show_bug.cgi?id=18928
Walter Bright changed:
What|Removed |Added
CC||bugzi...@digitalmars.com
--- Comment #2
https://issues.dlang.org/show_bug.cgi?id=18928
--- Comment #1 from Manu ---
I'm suspecting an issue with RVO.
I think the MSC ABI has changed to guarantee RVO at some point in the last few
years... but I'm not sure why the static call and virtual call appear to be
called using different calling
https://issues.dlang.org/show_bug.cgi?id=18928
Manu changed:
What|Removed |Added
Keywords||C++, industry
--
13 matches
Mail list logo
