- User Since
- Dec 28 2012, 2:34 PM (272 w, 5 d)
Did you consider writing a custom parser as was mentioned on the bug? It really seems like it would make the code better rather than going through these contortions to fit the parsing into llvm::cl.
Fri, Mar 16
Thu, Mar 15
Probably the ideal test would be an object file with three weak externals with each of the possible characteristics. Then test that the archive symbol table contains only an entry for the symbol with characteristics IMAGE_WEAK_EXTERN_SEARCH_ALIAS.
I am tempered to just document in the language reference that it is invalid for a pass to drop local_unnamed_addr. With that the linker job is even simpler, it doesn't need to check the LTO produced files, just canBeOmittedFromSymbolTable in very IR symbol table.
Peter, what do you think?
I'd be fine with rejecting.
I will commit when D44501 is approved :)
- Address review comments
Wed, Mar 14
- Add a test that makes sure that we don't merge string literals with different alignments
Do you have a test to make sure that a wide string literal doesn't point into a misaligned offset into a narrow string literal?
- Align MergeChunks correctly
- Check contents of .rdata
Tue, Mar 13
Fri, Mar 9
Thu, Mar 8
Wed, Mar 7
- Update tests and fix a couple of benign bugs in the code
Tue, Mar 6
If the backend will be changed so that it will not depend on IR type names
Why is this a driver flag? This seems like it ought to be a cc1-only flag to me.
I guess this can be abandoned now, right?
Mon, Mar 5
Tue, Feb 27
The code is ready for review, but the tests still need to be updated. Up to you whether you'd like to take a look now or wait for the tests.
@chandlerc ping -- let me know whether this MI-based implementation seems reasonable to you.
Mon, Feb 26
Fri, Feb 23
This seems like a reasonable fix to me, but I'd like to know what Teresa thinks.
Thu, Feb 22
Apologies for missing this patch. It seems to be fixing the same problem as D43132, but that patch also handles trampoline functions correctly, so I think we should go with that one.
Wed, Feb 21
My personal opinion is that bindings should be off by default because I doubt the majority of users need it.
Tue, Feb 20
- First draft of MI-based lowering -- tests to come
Feb 16 2018
LGTM with an updated commit message.
Isn't --full-shutdown necessary for the leak sanitizer to work correctly?
The unnamed_addr attribute means that the address is semantically significant
Feb 15 2018
- Remove unused file
There is an alternative which is to teach linker to deduct auto hide property from the bitcode file and overwrite the linkage type afterwards. For C API, there is a new scope LTO_SYMBOL_SCOPE_DEFAULT_CAN_BE_HIDDEN can be used to do that but that will be forced the work onto all the linker which is probably not desirable but it solves the same problem.
The last time this was attempted (D28978) we established that the linker can do this itself. I don't think the situation has changed since then.
Feb 14 2018
Maybe I'm missing something, but I don't see why we need this attribute. Couldn't clang have been changed to implement -ftrap-function by generating a call to the trap function instead of emitting an llvm.trap intrinsic call?
Feb 13 2018
Feb 12 2018
I have no strong feelings about what happens in the legacy API, so probably someone at Apple should take a look first.
LGTM with a few spelling/grammar nits