Description of problem: when trying to build ia32el with gcc that ships with rhel4 we hit some minor problems, which made the compilation fail. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
Created attachment 106317 [details] fixes the problem of compiling with new gcc
The ia32el srpms have this fixed already (by removing inline keyword where it should not be used as the function really is not inlinable). See ia32el-gcc34.patch patch.
we considered this solution and prefer to remove the -Winline switch. can you please reconsider it ? thanks, yoav.
Curious, why you think removing -Winline is better? IMHO it is quite useful warning, if you have code that you intend to be inlined, it helps you ensure it really is inlined. What purpose is for the inline keyword on functions that really can't be inlined (they don't have function bodies defined in the header that has the inline keywords nor anything else that is included in the .c files that are using these routines)?
the argument was that it depends on the compiler, whether a function can be inlined or not. and we don't want to get into a situation where we need to change the sources for each new version of gcc (or intel compiler when we'll start using it).
It seems the noaltroot patch was already removing -Winline. Anyway, I have added the patch as is and removed from noaltroot the same hunks (because on RHEL3 noaltroot patch is not applied).
Intel, closing this out, as it hasn't come up in recent discussions. Please reopen if the problem is still present.