[PATCH] configure: check for libstdc++ and zlib, dynamic and static
Bryan Hundven
bryanhundven@gmail.com
Sat Jan 22 16:59:00 GMT 2011
On Sat, Jan 22, 2011 at 8:47 AM, Yann E. MORIN
<yann.morin.1998@anciens.enib.fr> wrote:
> Bryan, All,
>
> On Saturday 22 January 2011 00:02:40 Bryan Hundven wrote:
>> configure: check for libstdc++ and zlib, dynamic and static
>>
>> Debian packages both static and dynamic libraries in their -dev package,
>> but Fedora packages them seperately.
>>
>> Normal builds need libstdc++.a, and static toolchains require libz.a to
>> be installed on the host.
>
> And do we need libstdc++ when doing a C-only toolchain? If not, then it
> would be a bit over-stressing to unconditionally require libstdc++.
>
> Well, IIRC, libstdc++ is pulled by one of the companion library (can't
> remember which, though), which comes to be required only with gcc-4.4.
> So if one builds a toolchain with gcc-4.3 or earlier, then libstdc++ is
> not required.
>
> Then, a run-time check like the patch by Anthony would be better, I
> think. And it should be conditional to requiring libstdc++. Something
> in scripts/build/cc/gcc.sh, like:
> [...]
> libstdcxx_required=blabla
> if [ "${libstdcxx_required}" = "y" ]; then
> # Same test as suggested by Anthony
> fi
> [...]
I think both a configure time warning and a runtime check would be
helpful. I would hate to have gotten part way through the build to
find out I was missing a static library on my build box.
Maybe If I got a warning during configuration saying:
"Warning: Some configurations require <some library> that seems to be
missing on your computer."
And use Tony's patch to check during runtime and error out.
> This is gcc-specific, so has to go into the gcc build script. The issue
> with that is the test will occur only after some time has elapsed. If we
> could do it earlier in the process, that'd be nice.
>
> Maybe we could generalise the do_libc_check_config, and add such a step
> to all components, to give them the oportunity to bail-out early if
> something is wrong, that we can not test at ./configure-time.
>
> Regards,
> Yann E. MORIN.
>
> --
> .-----------------.--------------------.------------------.--------------------.
> | Yann E. MORIN | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
> | +33 662 376 056 | Software Designer | \ / CAMPAIGN | ___ |
> | +33 223 225 172 `------------.-------: X AGAINST | \e/ There is no |
> | http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL | v conspiracy. |
> '------------------------------^-------^------------------^--------------------'
>
--
For unsubscribe information see http://sourceware.org/lists.html#faq
More information about the crossgcc
mailing list