Hi, Using a JIT engine on LLVM-C currently fails on SVN because the LLVMInitialize???Info is never called. The attached patch fixes this. BTW, would it be possible to include a simple LLVM-C test case, like http://npcontemplation.blogspot.com/2008/06/secret-of-llvm-c-bindings.html, into the test suite, and build/test regularly to prevent regressions like this? Jose -------------- next part -------------- A non-text attachment was scrubbed... Name: llvm-c-initialize-target-info.patch Type: text/x-diff Size: 1480 bytes Desc: not available URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20090815/7704b80c/attachment.patch>
Applied as r79307, thanks On Sat, Aug 15, 2009 at 8:57 AM, José Fonseca<jose.r.fonseca at gmail.com> wrote:> Hi, > > Using a JIT engine on LLVM-C currently fails on SVN because the > LLVMInitialize???Info is never called. > > The attached patch fixes this. > > BTW, would it be possible to include a simple LLVM-C test case, like > http://npcontemplation.blogspot.com/2008/06/secret-of-llvm-c-bindings.html, > into the test suite, and build/test regularly to prevent regressions > like this?Yes. This should probably be done as a unit tests, for example we already have some JIT unittests for the C++ API. Want to make a patch for it? - Daniel> Jose > > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev > >
Kenneth Uildriks
2009-Aug-21 14:48 UTC
[LLVMdev] Patch: Initialize target info for LLVM-C.
Does the 2.5 release version also fail for this reason? I'm trying to JIT (through LLVM-C) from code generated by my compiler, and it's segfaulting... I'd like to know if this is the problem or if something else is going wrong. On Mon, Aug 17, 2009 at 10:06 PM, Daniel Dunbar<daniel at zuster.org> wrote:> Applied as r79307, thanks > > On Sat, Aug 15, 2009 at 8:57 AM, José Fonseca<jose.r.fonseca at gmail.com> wrote: >> Hi, >> >> Using a JIT engine on LLVM-C currently fails on SVN because the >> LLVMInitialize???Info is never called. >> >> The attached patch fixes this. >> >> BTW, would it be possible to include a simple LLVM-C test case, like >> http://npcontemplation.blogspot.com/2008/06/secret-of-llvm-c-bindings.html, >> into the test suite, and build/test regularly to prevent regressions >> like this? > > Yes. This should probably be done as a unit tests, for example we > already have some JIT unittests for the C++ API. Want to make a patch > for it? > > - Daniel > >> Jose >> >> _______________________________________________ >> LLVM Developers mailing list >> LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu >> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev >> >> > > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev >
Possibly Parallel Threads
- [LLVMdev] Patch: Initialize target info for LLVM-C.
- [LLVMdev] Patch: Initialize target info for LLVM-C.
- [LLVMdev] [PATCH]: MSVC build enhancements
- [LLVMdev] Prevent unbounded memory consuption of long lived JIT processes
- [LLVMdev] Prevent unbounded memory consuption of long lived JIT processes