PR1317 says that it is resolved. Looking at the details of the report, it says that a fix had been committed for the "instant case." Am I right that means a fix has been submitted that makes LLVMHello work with opt, but that the general problem has not been resolved and is reported in PR1318? Reid Spencer wrote:> Hi Ryan, > > On Tue, 2007-04-10 at 00:12 -0500, Ryan M. Lefever wrote: > >>I checked out llvm from cvs & llvm-gcc from svn last night and again >>tonight. Each time they compiled and installed fine. After installing >>them, I recompiled compiler transforms I had written for opt. opt seems >>to load the my transform libraries fine, but it complains: >> >>opt: Unknown command line argument '-mytransform' > > >>whenever I try to specify one of my transforms on the opt command line. > > > Please see PR1317 and PR1318. I believe you're running into the same > issue. Briefly, the issue is that option processing has changed to not > be so oriented towards static initialization time. This has changed the > order of things and now your option is not recognized because it hasn't > been registered properly. > > >> I checked the names of the transforms and they are correct. In >>addition, I did a opt --help and my transforms were lisetd. > > > Yes, this isn't your bug. > > >>So, I switched back to a version of llvm & llvm-gcc from about a month >>ago, and opt recognizes and runs my transforms. Did something change >>with opt? > > > These are probably the result of changes to the CommandLine.cpp file in > the last few days. All LLVM tools use a common facility for processing > command line options. > > >>Did something change with the way transforms are registered >>with opt? > > > No, the registration mechanism is the same. > > Reid. > > > >>Regards, >>Ryan >> >>_______________________________________________ >>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-- Ryan M. Lefever [http://www.ews.uiuc.edu/~lefever]
On Tue, 2007-04-10 at 01:27 -0500, Ryan M. Lefever wrote:> PR1317 says that it is resolved. Looking at the details of the report, > it says that a fix had been committed for the "instant case." Am I > right that means a fix has been submitted that makes LLVMHello work with > opt, but that the general problem has not been resolved and is reported > in PR1318?Yes, that's pretty much it. 1317 was about LLVMHello. 1318 is a more substantial bug that has not been resolved yet. I mentioned both because perhaps the fix for LLVMHello (don't link LLVM libraries into the module) will work for you as well. Reid.> > > Reid Spencer wrote: > > Hi Ryan, > > > > On Tue, 2007-04-10 at 00:12 -0500, Ryan M. Lefever wrote: > > > >>I checked out llvm from cvs & llvm-gcc from svn last night and again > >>tonight. Each time they compiled and installed fine. After installing > >>them, I recompiled compiler transforms I had written for opt. opt seems > >>to load the my transform libraries fine, but it complains: > >> > >>opt: Unknown command line argument '-mytransform' > > > > > >>whenever I try to specify one of my transforms on the opt command line. > > > > > > Please see PR1317 and PR1318. I believe you're running into the same > > issue. Briefly, the issue is that option processing has changed to not > > be so oriented towards static initialization time. This has changed the > > order of things and now your option is not recognized because it hasn't > > been registered properly. > > > > > >> I checked the names of the transforms and they are correct. In > >>addition, I did a opt --help and my transforms were lisetd. > > > > > > Yes, this isn't your bug. > > > > > >>So, I switched back to a version of llvm & llvm-gcc from about a month > >>ago, and opt recognizes and runs my transforms. Did something change > >>with opt? > > > > > > These are probably the result of changes to the CommandLine.cpp file in > > the last few days. All LLVM tools use a common facility for processing > > command line options. > > > > > >>Did something change with the way transforms are registered > >>with opt? > > > > > > No, the registration mechanism is the same. > > > > Reid. > > > > > > > >>Regards, > >>Ryan > >> > >>_______________________________________________ > >>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 >
On Tue, 10 Apr 2007, Reid Spencer wrote:> Yes, that's pretty much it. 1317 was about LLVMHello. 1318 is a more > substantial bug that has not been resolved yet. I mentioned both > because perhaps the fix for LLVMHello (don't link LLVM libraries into > the module) will work for you as well.I will try to get this bug fixed today, -Chris>> >> >> Reid Spencer wrote: >>> Hi Ryan, >>> >>> On Tue, 2007-04-10 at 00:12 -0500, Ryan M. Lefever wrote: >>> >>>> I checked out llvm from cvs & llvm-gcc from svn last night and again >>>> tonight. Each time they compiled and installed fine. After installing >>>> them, I recompiled compiler transforms I had written for opt. opt seems >>>> to load the my transform libraries fine, but it complains: >>>> >>>> opt: Unknown command line argument '-mytransform' >>> >>> >>>> whenever I try to specify one of my transforms on the opt command line. >>> >>> >>> Please see PR1317 and PR1318. I believe you're running into the same >>> issue. Briefly, the issue is that option processing has changed to not >>> be so oriented towards static initialization time. This has changed the >>> order of things and now your option is not recognized because it hasn't >>> been registered properly. >>> >>> >>>> I checked the names of the transforms and they are correct. In >>>> addition, I did a opt --help and my transforms were lisetd. >>> >>> >>> Yes, this isn't your bug. >>> >>> >>>> So, I switched back to a version of llvm & llvm-gcc from about a month >>>> ago, and opt recognizes and runs my transforms. Did something change >>>> with opt? >>> >>> >>> These are probably the result of changes to the CommandLine.cpp file in >>> the last few days. All LLVM tools use a common facility for processing >>> command line options. >>> >>> >>>> Did something change with the way transforms are registered >>>> with opt? >>> >>> >>> No, the registration mechanism is the same. >>> >>> Reid. >>> >>> >>> >>>> Regards, >>>> Ryan >>>> >>>> _______________________________________________ >>>> 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 >> > > _______________________________________________ > LLVM Developers mailing list > LLVMdev at cs.uiuc.edu http://llvm.cs.uiuc.edu > http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev >-Chris -- http://nondot.org/sabre/ http://llvm.org/