[Bug tree-optimization/36439] [4.3 Regression] very long compile-time in PRE building gimp-plugin-registry

2009-03-14 Thread vapier at gentoo dot org


-- 

vapier at gentoo dot org changed:

   What|Removed |Added

 CC||toolchain at gentoo dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=36439

--- You are receiving this mail because: ---
You reported the bug, or are watching the reporter.


-- 
To UNSUBSCRIBE, email to debian-gcc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org



Re: [RFC] [Cross Toolchain] Multiarch and sysrooted toolchain

2009-03-14 Thread Goswin von Brederlow
Hector Oron  writes:

> Hello,
>
 , and there is generally no need to
 install anything but libraries and headers into /usr/ -- so I
 don't think there is a pressing need to replicate a filesystem hierarchy
 standard below a triplet directory.
>> 
>>> True, however, that is not a sufficient reason to not
>>> move /usr/ to /usr/lib/ and /usr/include/
>>> if it means getting such support into the core Debian packaging tools.
>> 
>> Indeed, however this makes building stuff without the packaging tools a lot
>> harder -- for example I need to patch gcc to recognize these paths if I
>> build gcc by hand. It might be a lot easier to have the packaging tools
>> handle the "current" layout than to patch all the software that assumes
>> that software is generally installed into "include" and "lib" dirs under a
>> common prefix (such as most GNU software that uses the autoconf macros to
>> find installed software).
>
> That's totally right and that is my point, I really think multiarch is
> not well designed to fit into cross compiling. They (dpkg-core) might
> want us to do extra work which will break our stuff.

If gcc looks into the tripplet dirs by default then cross-compiling
can just use the normal system directories. The sys-root would be / in
that case. The common prefix "".

> And as from the point of view of multiarch we probably have a nice,
> simple and working solution right now, without even notice it. The only
> reason I found against our approach is:
>
> "
> Why not prefix/arch-os/lib/ (and include/)?
> It would pollute the prefix directory. Can you imagine adding one entry
> for each target to the root and /usr directories? Better that they go
> under the prefix/lib/ (and include/) directories which already contain
> many files. "
>
> Which in practice it is not so bad to do all the extra work that
> multiarch needs to get done.
>
> Please correct me if I am wrong

You are wrong. :)

The toolchain does not yet look in all the right places. Neither for
the multiarch nor the corss-compile way of putting the prefix. It is
in a state where both ways are used and neither is complete enough for
a full system.

MfG
Goswin


-- 
To UNSUBSCRIBE, email to debian-gcc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org



Re: [RFC] [Cross Toolchain] Multiarch and sysrooted toolchain

2009-03-14 Thread Goswin von Brederlow
Simon Richter  writes:

> Hi,
>
>> >, since
>> > they have entirely different objectives
>
>> Not entirely different - the objective for the packaging tools is
>> actually the same, to have packages install cleanly without changes on
>> systems with a different architecture triplet.
>
> I'm not sure this can be achieved at all, as we still need a root
> filesystem that isn't prefixed with the architecture triplet.

Why? At least libraries are just fine in the triplet dirs. As said
binaries should not move. For cross-building dpkg-cross can move them
during unpacking if that is needed.

> The other issue is that generally, the 32/64 bit distinction does not
> necessarily mean that we use a different triplet. i386/amd64 does, at least
> in Debian, but that is optional as well and could be changed in the gcc
> package, which would give us a situation where only clearly incompatible
> arches need to be installed into a triplet directory.

Then change the triplets. Having binary files for multiple
architectures in the same triplet dirs works neither for multiarch nor
cross-building.

>> >, and there is generally no need to
>> > install anything but libraries and headers into /usr/ -- so I
>> > don't think there is a pressing need to replicate a filesystem hierarchy
>> > standard below a triplet directory.
>
>> True, however, that is not a sufficient reason to not
>> move /usr/ to /usr/lib/ and /usr/include/
>> if it means getting such support into the core Debian packaging tools.
>
> Indeed, however this makes building stuff without the packaging tools a lot
> harder -- for example I need to patch gcc to recognize these paths if I
> build gcc by hand. It might be a lot easier to have the packaging tools
> handle the "current" layout than to patch all the software that assumes
> that software is generally installed into "include" and "lib" dirs under a
> common prefix (such as most GNU software that uses the autoconf macros to
> find installed software).
>
>Simon

The current layout does not result in a bootable system. We do need

/arch-os-libc/lib/

and then stuff already needs patching.

Also current setup is:
% ldconfig -v | grep :
/usr/lib/i486-linux-gnu:
/usr/local/lib:
/lib/x86_64-linux-gnu:
/usr/lib/x86_64-linux-gnu:
/lib:
/lib32:
/usr/lib:
/usr/lib32:
/usr/lib32/i586: (hwcap: 0x0004)
/usr/lib32/i486: (hwcap: 0x0002)
/usr/lib32/i686: (hwcap: 0x0008)
/usr/lib32/i686/cmov: (hwcap: 0x00088000)

ldconfig has no idea about cross-compile dirs. And someone suggested
they should be system dirs hardcoded into the binary and not via
conffiles.

MfG
Goswin


-- 
To UNSUBSCRIBE, email to debian-gcc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org



Re: [RFC] [Cross Toolchain] Multiarch and sysrooted toolchain

2009-03-14 Thread Hector Oron
Hello,

> The toolchain does not yet look in all the right places. Neither for
> the multiarch nor the corss-compile way of putting the prefix. It is
> in a state where both ways are used and neither is complete enough for
> a full system.

So, would it be fine to send an addendum to multiarch[1] document
taking into account cross environments?

In case, we write up that file, preferred way for cross environments
would be a sysroot , where in multiarch do we fit our cross 
? Or we should just go for the splitted /* option (which won't be
useful until multiarch is ready) ?

>From cross toolchain point of view, i would suggest to stay with
upstream maintainers layout, which would require a change on
dpkg-cross, moving
/include into /usr/include, that way we would be able to export .

Regards

[1] http://lackof.org/taggart/hacking/multiarch/

-- 
 Héctor Orón


--
To UNSUBSCRIBE, email to debian-gcc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org