Re: [Tinycc-devel] configury (was const_wanted)

2011-08-03 Thread grischka
Thomas Preud'homme wrote: Why is it necessary to look in two directories at runtime for elf_interp and for the crt objects? Actually I realized only today how big mistake it was for the elf interpreter (it's part of the ABI). This will be fixed in Debian today or tomorrow. As to crt objects, I

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-03 Thread Thomas Preud'homme
Le mercredi 3 août 2011 19:36:47, grischka a écrit : > Thomas Preud'homme wrote: > >> Why is it necessary to look in two directories at runtime for > >> elf_interp and for the crt objects? > > > > Actually I realized only today how big mistake it was for the elf > > interpreter (it's part of the A

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-03 Thread grischka
Thomas Preud'homme wrote: Actually I'd like to be able to try 2 prefixs, to reduce the dependencies between glibc and tcc. This way tcc could work well both with the old glibc (which install crt objects in /usr/lib) and the current glibc (which install crt objects in /usr/lib/). If not OK for y

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-06 Thread Thomas Preud'homme
Le jeudi 4 août 2011 00:53:59, grischka a écrit : > Thomas Preud'homme wrote: > > Actually I'd like to be able to try 2 prefixs, to reduce the dependencies > > between glibc and tcc. This way tcc could work well both with the old > > glibc (which install crt objects in /usr/lib) and the current gli

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-06 Thread grischka
Thomas Preud'homme wrote: Does gcc try prefixes for crt objects? I don't know for the upstream gcc but I did a try on my debian system > and it is: [...] Sure, why else should they have 250kb of gcc.c driver code. Attached is a patch to implement the solution you were asking for. Well, n

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-06 Thread Thomas Preud'homme
Le samedi 6 août 2011 16:55:15, grischka a écrit : > Thomas Preud'homme wrote: > >> Does gcc try prefixes for crt objects? > > > > I don't know for the upstream gcc but I did a try on my debian system > > > > and it is: > > [...] > > Sure, why else should they have 250kb of gcc.c driver code. >

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-07 Thread grischka
Thomas Preud'homme wrote: Anyway I pushed some patches so you can have crtprefix1:crtprefix2. Note that it's also sysrooted. I don't know why you changed that. Did I? When? I was probably not careful enough. No, it seems you wanted to fix something. It's just not clear what: http://repo.

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-08 Thread Thomas Preud'homme
Le dimanche 7 août 2011 21:38:44, grischka a écrit : > Thomas Preud'homme wrote: > >> Anyway I pushed some patches so you can have crtprefix1:crtprefix2. > >> Note that it's also sysrooted. I don't know why you changed that. > > > > Did I? When? I was probably not careful enough. > > No, it seem

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-08 Thread grischka
Thomas Preud'homme wrote: So if the file system layout change it's quite normal that gcc have to handle these changes. Why is that normal, quite? We just proved that tcc can work for multiarch without any encoded knowledge at all what multiarch is. ./configure --prefix=/home/robotux/local -

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-12 Thread Thomas Preud'homme
Le lundi 8 août 2011 18:03:16, grischka a écrit : [SNIP] > > Oops! If this is b/include with 8 backslashes, then > I think we must use something else. > > What about ?B as in "?B/include"? Or maybe "{B}/include"? cstr_new(&str); for (p = in; c = *p, c != '\0' && c !

Re: [Tinycc-devel] configury (was const_wanted)

2011-08-12 Thread grischka
Thomas Preud'homme wrote: +if (c == '{' && p[1] && p[2] == '}') { +c = p[1], p += 2; +if (c == 'B') +cstr_cat(&str, s->tcc_lib_path); I guess you did it this way to avoid just forbidding {b} but allowing any {x} with x ≠ b. Not