Kevin O'Connor escreveu: > On Mon, Nov 20, 2006 at 02:27:59PM +0000, Pedro Alves wrote: > >> Pedro Alves escreveu: >> >>> Kevin O'Connor escreveu: >>> >>> >>>> On Mon, Nov 20, 2006 at 01:18:21AM +0000, Pedro Alves wrote: >>>> >>>> >>>> >>>>> I have never seen a WinCE < 3 machine, and I don't really know for sure >>>>> what may be different there. >>>>> Currently all the toolchain is built with CE 3 as the default (linker >>>>> and headers). >>>>> >>>>> >>>>> >>>> I was looking at this a little further. It seems the mingw code is >>>> calling GetCommandLine. However, that function was introduced in CE3: >>>> > > Hi Pedro, > > I have confirmation that the app works on the WinCE 2.1 strongarm > processor if we bypass the WinMainCRTStartup and explicitly call > __gccmain from WinMain. > > As to a fix - I'm a bit confused by the current mingw crt1.c code. > Why do we have WinMainCRTStartup call main() which then calls WinMain. > Wouldn't it be simpler to have a dummy WinMain that calls main(). > This way, the code doesn't have to scan all the command-line arguments > if the application isn't interested in them. > > Yes, I also think it would be simpler. But, it requires a gcc change (simple one, I think). Currently the __gccmain call is emitted on main, and we would need to change it to be emitted on WinMain, because then a main may not even exist. I have been thinking about this for a few days, and as a coincidence I am fixing a pair of bugs in gcc and gdb that somewhat touches these fields: (Actually I have the fixes, I just need the harddrive space to bootstrap/regtest gcc :) ) http://sources.redhat.com/ml/gdb/2006-11/msg00134.html
Before doing the main/WinMain WinMain/main change, I want to ask the MinGW guys why they didn't do it that way in the first place, as we may be missing something obvious. > Also, what are the requirements in these files for 386 compatibility? > It seems like there isn't that much overlap between between desktop > windows and ce windows. In overview there is :) Well the steps are the same, but yes, a lot if #ifdef'ed out. Fortunately, this is code that we shouldn't need to look at much. :) > I find the ifdefs to be hard to follow - in > particular because there are a number of ifdefs (MSV_CRT, UNDER_CE, > COREDLL) and it isn't clear which are actually in effect on CE. I > wonder if it be simpler to have a ce specific crt1.c file. > > In stock MinGW, they support two different MSFT runtimes. msvcrt.dll and crtdll.dll. We added a third: coredll.dll. That's where the __MSVCRT__, __CRTDLL__ and __COREDLL__ come from. The UNDER_CE ones I guess could be changed into __COREDLL__. > Finally, is this something you're willing to fix? > > Sure, but probably only in a week unfortunately. Cheers, Pedro Alves ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Cegcc-devel mailing list Cegcc-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/cegcc-devel
