X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f X-Recipient: djgpp AT delorie DOT com Message-ID: Content-Type: multipart/alternative; boundary="_7baa6252-fb8a-4968-ba45-3dbf23d99a1d_" X-Originating-IP: [71.112.20.48] From: Jay To: DJ Delorie CC: Subject: RE: libstdc++ writev/2.04/patches upstream? Date: Mon, 7 Jul 2008 21:31:24 +0000 Importance: Low In-Reply-To: <200807071920.m67JKA4v032518@envy.delorie.com> References: <200807070405 DOT m67451dZ010910 AT delorie DOT com> <200807071920 DOT m67JKA4v032518 AT envy DOT delorie DOT com> MIME-Version: 1.0 X-OriginalArrivalTime: 07 Jul 2008 21:31:24.0860 (UTC) FILETIME=[CEB317C0:01C8E078] Errors-To: nobody AT delorie DOT com X-Mailing-List: djgpp AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk --_7baa6252-fb8a-4968-ba45-3dbf23d99a1d_ Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable I'll have to gradually work through them. Thanks.> That might mean it's har= dcoded for the newer features. Automated> testing for djgpp features is tri= cky at best. Why is it any more difficult than any other crossed situation (of various s= orts)? I've been going at this cross build stuff a few weeks now with some success= =2C still a bit more to do=2C and I understand/believe that 1) you must hav= e the "sys-root" and possibly 2) cross binutils before cross gcc (I've been= going in that order -- e.g. so you can build stuff like libgcc). People mi= ght try to get by "without headers" or with only "sys-include"=2C and those= situations could fall back=2C but if you have sys-root and cross binutils= =2C you should at least be able to compile and link C test cases when confi= guring/building libstdc++ (and compile C++ test cases). =20 (sys-root / build-sysroot / with-headers /with-libs is a bit confusing=2C c= ould be friendly imho) =20 Thanks=2C - Jay> Date: Mon=2C 7 Jul 2008 15:20:10 -0400> From: dj AT delorie DOT com> To: j= ayk123 AT hotmail DOT com> CC: djgpp AT delorie DOT com> Subject: Re: libstdc++ writev/2.= 04/patches upstream?> > > > What is the status of 2.04?> > If we can get th= e people who know about 2.04-related unfixed bugs to> summarize=2C maybe we= can just release it as-is and put off those fixes> for a quick 2.05.> > > = Should be easier to start with a "release" and move to "current" ?> > If yo= u're interested in helping=2C please start with cvs.> > > What is the statu= s of getting diffs propagated upstream?> > For djgpp=2C post them to djgpp-= workers AT delorie DOT com> > > But gcc the patches don't appear present in 4.3.1.= > > I checked just one or two.> > I understand there is no obligation to su= bmit or accept diffs.> > But it would be nice.> > gcc-patches AT gcc DOT gnu DOT org> = > but you need to get the OK of the original patch author=2C and find out> = why they haven't been contributed yet - there may be djgpp-specific> reason= s to keep it out of the general sources.> > > libstdc++-v3 fails due to __i= ov being incomplete.> > It has something to do with > > GLIBCXX_CHECK_WRITE= V> > vs.> > > > > > *djgpp)> > ...> > D:\djgpp\gnu\gcc-4.23\libstdc++-v3\cr= ossconfig.m4(61): AC_DEFINE(HAVE_WRITEV)> > > > > > crossconfig.m4 is hardc= oded for 2.04?> > I see current cvs has writev=2C but that 2.03 does not.> = > That might mean it's hardcoded for the newer features. Automated> testing= for djgpp features is tricky at best.> > > I suspect if I setup an LFN-awa= re native DJGPP=2C this would configure> > correctly=2C but that cross buil= ds avoid the probe=2C but that probing> > would work=2C since by now there = is a cross-ld and I do get a gcc.exe.> > You could try a native djgpp build= =2C assuming you have LFN.= --_7baa6252-fb8a-4968-ba45-3dbf23d99a1d_ Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable I'll have to gradually work through them. Thanks.

>=3B That might = mean it's hardcoded for the newer features. Automated
>=3B testing for= djgpp features is tricky at best.

Why is it any more difficult than any other crossed situation (of various s= orts)?
I've been going at this cross build stuff a few weeks now with some success= =2C still a bit more to do=2C =3Band I understand/believe that 1) you m= ust have the "sys-root" and possibly 2) cross binutils before cross gcc (I'= ve been going in that order -- e.g. so you can build stuff like libgcc). Pe= ople might try to get by "without headers" or with only "sys-include"=2C an= d those situations could fall back=2C but if you have sys-root and cross bi= nutils=2C you should at least be able to compile and link C test =3Bcas= es when configuring/building libstdc++ (and compile C++ test cases).
 =3B
(sys-root / build-sysroot / with-headers /with-libs is a bit confusing=2C c= ould be friendly imho)
 =3B
Thanks=2C
 =3B- Jay

>=3B Date: Mon=2C 7 Jul 2008 15:20:10 -0400
>= =3B From: dj AT delorie DOT com
>=3B To: jayk123 AT hotmail DOT com
>=3B CC: dj= gpp AT delorie DOT com
>=3B Subject: Re: libstdc++ writev/2.04/patches upstre= am?
>=3B
>=3B
>=3B >=3B What is the status of 2.04?
&= gt=3B
>=3B If we can get the people who know about 2.04-related unfix= ed bugs to
>=3B summarize=2C maybe we can just release it as-is and pu= t off those fixes
>=3B for a quick 2.05.
>=3B
>=3B >=3B S= hould be easier to start with a "release" and move to "current" ?
>=3B=
>=3B If you're interested in helping=2C please start with cvs.
&g= t=3B
>=3B >=3B What is the status of getting diffs propagated upstr= eam?
>=3B
>=3B For djgpp=2C post them to djgpp-workers AT delorie DOT c= om
>=3B
>=3B >=3B But gcc the patches don't appear present in = 4.3.1.
>=3B >=3B I checked just one or two.
>=3B >=3B I under= stand there is no obligation to submit or accept diffs.
>=3B >=3B Bu= t it would be nice.
>=3B
>=3B gcc-patches AT gcc DOT gnu DOT org
>=3B =
>=3B but you need to get the OK of the original patch author=2C and f= ind out
>=3B why they haven't been contributed yet - there may be djgp= p-specific
>=3B reasons to keep it out of the general sources.
>= =3B
>=3B >=3B libstdc++-v3 fails due to __iov being incomplete.
= >=3B >=3B It has something to do with
>=3B >=3B GLIBCXX_CHECK_W= RITEV
>=3B >=3B vs.
>=3B >=3B
>=3B >=3B
>=3B &g= t=3B *djgpp)
>=3B >=3B ...
>=3B >=3B D:\djgpp\gnu\gcc-4.23\li= bstdc++-v3\crossconfig.m4(61): AC_DEFINE(HAVE_WRITEV)
>=3B >=3B
= >=3B >=3B
>=3B >=3B crossconfig.m4 is hardcoded for 2.04?
&g= t=3B >=3B I see current cvs has writev=2C but that 2.03 does not.
>= =3B
>=3B That might mean it's hardcoded for the newer features. Autom= ated
>=3B testing for djgpp features is tricky at best.
>=3B
= >=3B >=3B I suspect if I setup an LFN-aware native DJGPP=2C this would = configure
>=3B >=3B correctly=2C but that cross builds avoid the pro= be=2C but that probing
>=3B >=3B would work=2C since by now there is= a cross-ld and I do get a gcc.exe.
>=3B
>=3B You could try a na= tive djgpp build=2C assuming you have LFN.

= --_7baa6252-fb8a-4968-ba45-3dbf23d99a1d_--