On Sat, 28 Mar 2009 20:02:05 +0100, Danny Backx <danny.ba...@scarlet.be> wrote: > On Sat, 2009-03-28 at 17:57 +0100, Vincent Torri wrote: >> On Sat, 28 Mar 2009, Danny Backx wrote: >> > You'd also need : >> > - binutils (but that's already in good shape) >> > - include files (IMO no distribution is good enough) >> > - libraries to approach the CE DLLs (e.g. our libcoredll.a) >> >> another thing: another Vincent proposed to move libce to mingw-w64 (there >> >> are already libw32 and libw64 there). What do you think of that ? > > We've made quite a few improvements - most CE-specific - to the include > files that we're using. I haven't seen another distribution that has this. > > So moving to another base would mean merging all that work. > > Before doing that, I'd like to see added value. Right now, I don't know of > any. > > Danny
Hello, I would like to react about about cegcc and development process. First these last weeks end I played with cegcc-4.4 and and to be able to generate it I first tried to report cegcc modifications but I had some issues until Pedro published another patch a few months ago that was working fine. The problem is during these month it was hard to progress and to get support, Pedro is very busy and I didn't see a message from him even when I was asking a question. So I tried to get some help from gcc or gcc-help but since we are modifying vanilla gcc, gcc folks couldn't help me. That's the problem of maintaining something outside gcc trunk, what are we suppose to do when we need some help but nobody answer ? What I don't like is the fact that everything depends on one man and if this person is busy we are stuck and we cannot improve things. While I am talking about cegcc-4.4 you should know that atcually there is NO INTEREST to use it because code is BIGGER and SLOWER that the current cegcc(at least for one application I have tested) and to me cegcc-4.4 was only an exercise to follow gcc trunk and later to be able to generate a cegcc-4.5 when it is released. Indeed it's much easier to report modifications from a recent gcc that from an old one. Another example is the cegcc bug, this is something I would like to see fixed and that's why I contacted a developper named Kai Tietz and who is the maintainer of mingw-w64. He had already helped me to fix a bug in windres and he is working hard to improve windows support. For instance he has implemented unicode crt for mingw-w64, he has fixed some issues with pe format (IAT was null), ... Since he is very active I find it logical to make a mingw-wce and I have started from scratch to declare wince includes, for now I don't have time to keep on working on it but I hope I will find some time later. The thing I would like is to be able to follow gcc/mingw-w64 new features, to be able to test and report bugs and not to wait every six months or every year. >From what I know gcc will add support for Linked Time Optimizations (LTO) very soon and this is something I would like to test to see how it can improve generated code. Let's say I find a bug with cegcc, I wouldn't be able to report it because to be able to reproduce it gcc developpers will have to patch their sources and this case we cannot talk about gcc anymore, I don't think they want to help people doing their modifications on their side. Vincent R. ------------------------------------------------------------------------------ _______________________________________________ Cegcc-devel mailing list Cegcc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/cegcc-devel
