Seen this? >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> >> make -C ./antlr all >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> make[1]: Entering directory `/usr/src/build/683384-ppc64/BUILD/build-frysk-0.0.1.2006.01.13/antlr-2.7.5/antlr' basename: too many arguments Try `basename --help' for more information. *** compiling 209 Java file(s) gcj4: unrecognized option '-sourcepath' actions/java/ActionLexer.java:31: error: Can't find default package 'antlr'. Check the CLASSPATH environment variable and the access to the archives /usr/src/build/683384-ppc64/BUILD/antlr-2.7.5/antlr/debug/DebuggingCharScanner.java:4: error: Package 'antlr.collections' not found in import. import antlr.collections.*; ^ 2 errors xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx >> E R R O R << ============================================================ /usr/bin/gcj4 -C -mminimal-toc -d . -sourcepath /usr/src/build/683384-ppc64/BUILD/antlr-2.7.5 -classpath /usr/src/build/683384-ppc64/BUILD/build-frysk-0.0.1.2006.01.13/antlr-2.7.5/antlr/antlr.jar Version.java perhaphs it's fixed in a newer antlr?
I didn't see this on ppc64 with the old source rpms. But while I try to build the latest one, it didn't got that far. It stoped at an earlier place... Will try to see whether I can work around them. BTW, did you run the test on a ppc64 machine? I am just curious about how did you get that far. :-)
(In reply to comment #1) > I didn't see this on ppc64 with the old source rpms. But while I try to build > the latest one, it didn't got that far. It stoped at an earlier place... Will > try to see whether I can work around them. Perhaphs try building just antlr? The bug will likely either be me botching the config.* update, or a bug in the config stuff. frysk-imports/antlr should contain 2.7.5 antlr with patches applied. The RPM antlr patch created using: cvs -d :ext:sources.redhat.com:/cvs/frysk rdiff \ -u \ -r ANTLR_2_7_5 \ -r HEAD \ frysk-imports/antlr \ > antlr.patch You could also create a bug for the other problem and hang that off of 2188 (this bug's blocker). > BTW, did you run the test on a ppc64 machine? I am just curious about how did > you get that far. :-) I got curious and tried pushing the RPM through a PPC-64 RHEL 4 U3 build machine :-)
I think I found the reason. It is all the fault of basename. [root@localhost antlr-2.7.5.new]# basename /usr/bin/gcj4 -mminimal-toc -C basename: too many arguments Try `basename --help' for more information. Building it on x86 don't require "-mminimial-toc", it will output this: [root@localhost antlr-2.7.5.new]# basename /usr/bin/gcj4 -C gcj4 Add a quotation mark around will fix this: [root@localhost antlr-2.7.5.new]# basename "/usr/bin/gcj4 -mminimal-toc -C" gcj4 -mminimal-toc -C Can I get the write access of the CVS, then I can commit that myself?
2006-01-26 Andrew Cagney <cagney@redhat.com> Fix from Wu Zhou <woodzltc@cn.ibm.com>: * configure.in: Correctly quote basename applied to commands. * scripts/cpp.sh.in: Ditto. * scripts/cxx.sh.in: Ditto. * scripts/jar.sh.in: Ditto. * scripts/javac.sh.in: Ditto. * scripts/lib.sh.in: Ditto. * scripts/link.sh.in: Ditto.
* Fri Jan 27 2006 Andrew Cagney <cagney@redhat.com> 0.0.1.2006.01.22-0.EL4.5 - Update antlr.patch, update includes Wu Zhou's basename fix, and up-to-date config.guess and config.sub.