From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ian Lance Taylor To: egcs@cygnus.com Subject: Re: maintainer mode [was: Re: Building of generated parser files] Date: Wed, 27 Aug 1997 16:44:42 -0000 Message-id: <199708271526.LAA03697@subrogation.cygnus.com> In-reply-to: maintainer mode [was: Re: Building of generated parser files] X-SW-Source: 1997-08/0615.html Date: 27 Aug 1997 14:08:27 +0200 From: Samuel Tardieu >>>>> "Ian" == Ian Lance Taylor writes: Ian> I guess the main difficulty I see is the lack of a standard Ian> checksum program. Everything else seems workable. Why not including a small one in the distribution, to be built soon enough? That would lead us into a whole new set of problems, involving finding a working compiler for the build system which can generate programs we can execute. Remember, we might be building with a cross compiler. These problems are probably solvable, but at some point the solution becomes worse than the disease. Ian From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeffrey A Law To: egcs@cygnus.com Subject: Re: A config patch Date: Wed, 27 Aug 1997 16:44:42 -0000 Message-ID: <199708271526.LAA03697@subrogation.cygnus.com> In-reply-to: A config patch X-SW-Source: 1997-08/0617.html Message-ID: <19970827164442.tTsSb5SWxPl6PWcynXcKTz-oB68FN9boW_sgvl9k-oo@z> In message you write: > > so that [ -z "$PICFLAG" ] succeeds. But you would need to use > > libtool to think about getting OSF/1 shared libraries built -- > > gcc -shared certainly doesn't work. > > > > That may be true. But at least you can make libstdc++.so for > alpha/ELF. The thing you need to explain is _why_ this particular patch enables you to build a shared libstdc++. I don't have the time to grope around half a dozen makefiles and configure fragments to figure this out. You've obviously done so, so please save everyone some time by providing this information. And you also need to tell us why/how this library is correct if "gcc -shared" doesn't do the right thing on the alpha target. (IMHO) Building a broken libstdc++.so is worse than not building it at all. Jeff From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeffrey A Law To: egcs@cygnus.com Subject: Re: A config patch Date: Wed, 27 Aug 1997 16:44:42 -0000 Message-ID: <199708271526.LAA03697@subrogation.cygnus.com> In-reply-to: A config patch X-SW-Source: 1997-08/0616.html Message-ID: <19970827164442.We8eTv_7VnQ2rNU6p0Ort8R0hRDN2NAeGIQitxljIb8@z> In message <199708270841.BAA10195@rtl.cygnus.com>you write: > > There's no reference to these files in the toplevel configure.in; > > therefore, they will never be used. > > I forget the mechanism, but they get pulled via --enable-shared > in to attempt to build eg libiberty shared. Oh, I see it. You learn something new every day. What I still don't see is the point behind setting PICFLAG on this target -- yes, the if [-z "$PICFLAG"] test will now succeed, but, least least for libiberty, I don't see benefit. jeff