Mailing lists / fpc-devel / 2015-04
2 messages · 2 people ·
Reply by email Subscribe to this list Your subscription
Read-only archive. Use j and k to move between messages.
I am really happy to see this work moving forward. I think it is a great direction for the project.
Unfortunately, though, revision 30351 (the first point at which the llvm branch was reintroduced) has broken my code on Windows 64. Things are working fine though on 32 bit Windows and Linux.
I am in the process of trying to track this down now (and if I can I will report it as a bug), but in the meantime I was wondering if anybody else had experienced a similar issue?
Louis Salkind wrote on Wed, 08 Apr 2015:
I am really happy to see this work moving forward. I think it is a great direction for the project.
Direction is probably not the right word. It's no more a direction for the FPC project than adding support for AArch64, MSDOS or extended RTTI are. It's just an extra, optional, feature. And given the LLVM project's focus on features over backward compatibility, a too tight coupling between Clang and LLVM (formally they're separate libraries, but in practice implementation details leak through all over the place) and that they release way more often than we do, you'll probably have to stick to a select few LLVM versions when using it with FPC. Just to say that it won't be a miracle solution, and that it comes with its own share of downsides and limitations.
I've seen you've filed the bug report in the mean time. It's probably an alignment issue with the data emitted for widestrings. I'll look into it some time this week.
Jonas