This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC] Delete unused functions?

Secondly, and obviously I can't repeat this enough, all patch
contributions should report the target triples that they were
bootstrapped and regression tested on, and include the results of
such testing.

This is only an RFC for a patch I prepared in ten minutes, it's as far as possible from requesting approval. First of all, removing them may in turn make other functions unreachableThe patch as is bootstrapped.

Second, my script provides a quick start and all the functions suggested by my script should be hand-checked before removing them: I did this work on a tree without Ada to save checkout time, of course I would do the tests on a full tree, and on tree-ssa-pre-merge as well, before suggesting a serious patch. All I wanted was comments, and Andrew, Jan and you provided them indeed.

Being harsh is your maintainer's duty but, believe it or not (apart from, I admit, omitting the fact that I've bootstrapped/regtested the patch i686-pc-linux-gnu, which I *always* do), I do follow the rules in contribute.html. I'm sorry for the times I broke other platforms than this one, but I feel I've always fixed things properly. Unluckily I have no SPARC, IA64, Alpha or even x86-64 machines to test my changes on; I hope that I am still helping GCC in some way.

I did make a major mistake with the LTGT_EXPR patch; I have a better fix than what is currently on HEAD, and I'm trying to see why my machine does not run the Java tests, before submitting it.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]