This is the mail archive of the gcc-patches@gcc.gnu.org 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: IPA merge part 1: localize SSA variables


> Jan Hubicka wrote on 11/15/06 16:40:
> 
> >Maybe ;) but I am getting somewhat irritated by my own cgraph_ and
> >especially cgraph_varpool prefixes. Especially the second IMO just makes
> >code less readable by making it longer.  My plan is to rename
> >craph_varpool into varpool as variables are not cgraph after all and I
> >am not sure what I was thinking about when I introduced it at first
> >place.
> >Also I hope to have more clear split in between IPA infrastructure (ipa_
> >stuff), pass management and the cgraph datastructure itself that is all
> >somewhat intermixed right now in complettely misnamed cgraphunit.c...
> >
> Well, but this is different, I think.  For the cgraph functions you have 
> a single name space, but with the ILs we have several name spaces that 
> need to use the same terminology.  We are also bound by a troglodytic 
> implementation language, so there's little we can do.

This was my intention with cgraph/varpool too - to have two
datastructures with part of API symmetric (aka
cgraph_finalize/varpool_finalize, way to walk them and such) as for
large part of frontend functions and variables are the same, for most of
middle end the opposite is true.

But I guess we are again in agreement (I find myself writting GET_CODE
on trees and TREE_CODE on rtl quite often, symmetry would help)

Honza


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