https://bugzilla.redhat.com/show_bug.cgi?id=984605 --- Comment #19 from marcindulak <Marcin.Dulak@xxxxxxxxx> --- (In reply to Antonio Trande from comment #18) > (In reply to marcindulak from comment #17) > > This is also visible from all tests suspiciously failed in koji's build.log. > > There are ~1000 and ~4000 source files in nwchem compiled with debug and > > optimization flags, respectively, and the build time 2-3 hours. > > Have you even informed upstream ? i have been involved in some discussions on the nwchem forums, and the focus there is usually to get the code to compile and run at all on different high performance systems. It's a scientific code, and as such has only a "best effort" support. Also one cannot really rely on the QA test set (i don't know if the situation described here improved http://www.nwchem-sw.org/index.php/Special:AWCforum/st/id133/). The full test set includes ~400 cases (there is no concept of unit testing there - these are regular nwchem runs, some VERY heavy), and what i run during the rpmbuild is a small subset of 20. Of course when the nwchem RPM gets into Fedora i will drop a note at the nwchem forum, and mention compilation flags requirement problem. > > > > > Therefore, in my opinion, we have to stick to the default nwchem flags. > > I understand your position. I can just suggest to ask a Bundled Libraries > exception in order to resolve this problem. > http://fedoraproject.org/wiki/Packaging:No_Bundled_Libraries#Exceptions i was not clear enough. nwchem does not distribute GA, it compiles it in the executables. So - it's only a question about dropping %optflags or not. -- You are receiving this mail because: You are on the CC list for the bug. You are always notified about changes to this product and component _______________________________________________ package-review mailing list package-review@xxxxxxxxxxxxxxxxxxxxxxx https://admin.fedoraproject.org/mailman/listinfo/package-review