Jeff Trawick | 1 Apr 2003 01:11

Re: httpd 2.0.45-alpha on Solaris 9 w/Sun C 5.5 EA2 2003/01/09

Spinka, Kristofer wrote:
> Configure Used:
> ---------------
> CC=cc CFLAGS=-xO5 ./configure --with-mpm=worker --enable-rewrite
> --enable-so --enable-headers --enable-deflate
> 
> 
> Compiler Version:
> -----------------
> Sun ONE Studio 8 Compilers (Beta)
> Sun C 5.5 EA2 2003/01/09
> 
> 
> Proper libpthread.so Linking:

any idea why you get libpthread linked without hacking anything to refer 
to libpthread directly, but I don't?

> -----------------------------
> kspinka <at> stats1:/usr/local/apache2/bin] ldd ./httpd
>         libz.so.1 =>     /usr/lib/libz.so.1
>         libaprutil-0.so.0 =>     /usr/local/apache2/lib/libaprutil-0.so.0
>         libexpat.so.0 =>         /usr/local/apache2/lib/libexpat.so.0
>         libapr-0.so.0 =>         /usr/local/apache2/lib/libapr-0.so.0
>         libsendfile.so.1 =>      /usr/lib/libsendfile.so.1
>         librt.so.1 =>    /usr/lib/librt.so.1
>         libm.so.1 =>     /usr/lib/libm.so.1
>         libsocket.so.1 =>        /usr/lib/libsocket.so.1
>         libnsl.so.1 =>   /usr/lib/libnsl.so.1
>         libresolv.so.2 =>        /usr/lib/libresolv.so.2
(Continue reading)

William A. Rowe, Jr. | 1 Apr 2003 01:14

Re: mixing c & c++ on HP-UX

Andy, there were new fixes in 2.0.45 that should address this.  See...

http://cvs.apache.org/viewcvs.cgi/apr/dso/unix/dso.c

Bill

At 04:57 PM 3/31/2003, Andy Cutright wrote:
>hi all,
>
>there was some discussion on this list about HP-UX and c++ modules. we had a mixture like this running with
2.0.43. we've recently upgraded to 2.0.44, and we're now experiencing problems at load time. i'm
wondering if anyone is running 2.0.44 on HP-UX with c++ modules successfully,
>
>cheers,
>andy

Spinka, Kristofer | 1 Apr 2003 01:23

Re: httpd 2.0.45-alpha on Solaris 9 w/Sun C 5.5 EA2 2003/01/09

> any idea why you get libpthread linked without hacking anything to refer
> to libpthread directly, but I don't?

Did you remove the "-mt" from srclib/apr/configure /
srclib/apr/apr_threads.m4?

  /kristofer

Jeff Trawick | 1 Apr 2003 01:46

Re: httpd 2.0.45-alpha on Solaris 9 w/Sun C 5.5 EA2 2003/01/09

Spinka, Kristofer wrote:
>>any idea why you get libpthread linked without hacking anything to refer
>>to libpthread directly, but I don't?
> 
> 
> Did you remove the "-mt" from srclib/apr/configure /
> srclib/apr/apr_threads.m4?

I'm in the middle of trying that, but I'm far enough along that I can 
see that it works now :)

I don't see any indication that "-mt" in apr_threads.m4 works with 
anything but Solaris cc

http://cvs.apache.org/viewcvs.cgi/apr/Attic/threads.m4.diff?r1=1.9&r2=1.10&diff_format=h

that doesn't mean that some other unknown compiler on an obscure 
platform isn't relying on that code, but we're just going to have to 
find out :)

(i.e., delete the -mt check, which allows configure to discover the need 
for -lpthread)

Thanks,

Jeff

Spinka, Kristofer | 1 Apr 2003 01:56

Re: httpd 2.0.45-alpha on Solaris 9 w/Sun C 5.5 EA2 2003/01/09

  Just to confirm, here is a SPARC version of that ldd with "-mt" removed:

kspinka <at> build1:~/src/httpd-2.0.45] ldd .libs/httpd
        libz.so =>       /usr/local/lib/libz.so
        libaprutil-0.so.0 =>     /usr/local/apache2/lib/libaprutil-0.so.0
        libexpat.so.0 =>         /usr/local/lib/libexpat.so.0
        libapr-0.so.0 =>         /usr/local/apache2/lib/libapr-0.so.0
        libsendfile.so.1 =>      /usr/lib/libsendfile.so.1
        librt.so.1 =>    /usr/lib/librt.so.1
        libm.so.1 =>     /usr/lib/libm.so.1
        libsocket.so.1 =>        /usr/lib/libsocket.so.1
        libnsl.so.1 =>   /usr/lib/libnsl.so.1
        libresolv.so.2 =>        /usr/lib/libresolv.so.2
        libpthread.so.1 =>       /usr/lib/libpthread.so.1
        libdl.so.1 =>    /usr/lib/libdl.so.1
        libc.so.1 =>     /usr/lib/libc.so.1
        libgcc_s.so.1 =>         /usr/local/lib/libgcc_s.so.1
        libaio.so.1 =>   /usr/lib/libaio.so.1
        libmd5.so.1 =>   /usr/lib/libmd5.so.1
        libmp.so.2 =>    /usr/lib/libmp.so.2
        libthread.so.1 =>        /usr/lib/libthread.so.1
        /usr/platform/SUNW,UltraAX-i2/lib/libc_psr.so.1
        /usr/platform/SUNW,UltraAX-i2/lib/libmd5_psr.so.1

  /kristofer

On Mon, 31 Mar 2003, Jeff Trawick wrote:

> Spinka, Kristofer wrote:
> >>any idea why you get libpthread linked without hacking anything to refer
(Continue reading)

Jeff Trawick | 1 Apr 2003 02:19

Re: httpd 2.0.45-alpha on Solaris 9 w/Sun C 5.5 EA2 2003/01/09

Spinka, Kristofer wrote:
>  Builds and runs wonderfully on x86 and SPARC.
> 
>  I have not had the time for extensive testing, but just some simple
> feedback.
> 
> 
> Notes:
> ------
> I removed "-mt" from pthreads_cflags.

in case you don't watch the commit logs, your change to apr_thread.m4 
was just committed...  thanks so much for your help!

Sander Temme | 1 Apr 2003 02:27

Re: Apache 2.0.45 -alpha tarball candidates for testing

> www.apache.org server, but we need a variety of platforms and different
> configurations to have complete confidence in the release.

My Powerbook:

Darwin monalisa.sfo.covalent.net 6.4 Darwin Kernel Version 6.4: Wed Jan 29
18:50:42 PST 2003; root:xnu/xnu-344.26.obj~1/RELEASE_PPC  Power Macintosh
powerpc

With gcc:

[sctemme <at> MonaLisa] httpd-2.0.45 $ cc -v
Reading specs from /usr/libexec/gcc/darwin/ppc/3.1/specs
Thread model: posix
Apple Computer, Inc. GCC version 1173, based on gcc version 3.1 20020420
(prerelease)

unpacks, builds with bundled libtool and runs the perl-framework against
both worker and prefork. The only test that fails is that ssl/http one that
gives us a response without headers. We know about that.

My +1 (all +2 cents worth) for release.

S.

--

-- 
Covalent Technologies                             sctemme <at> covalent.net
Engineering group                                Voice: (415) 856 4214
303 Second Street #375 South                       Fax: (415) 856 4210
San Francisco CA 94107
(Continue reading)

Andy Cutright | 1 Apr 2003 02:36
Favicon

Re: mixing c & c++ on HP-UX

hi bill,

thanks for your help. that doesn't seem to be the root cause. we've been loading our modules without BIND_NOSTART for a while now on HP-UX. the program is dumping core in apr_pools.c:1976 with a SIGSEGV. i'm gonna start digging deeper,

cheers,
andy

William A. Rowe, Jr. wrote:
Andy, there were new fixes in 2.0.45 that should address this. See... http://cvs.apache.org/viewcvs.cgi/apr/dso/unix/dso.c Bill At 04:57 PM 3/31/2003, Andy Cutright wrote:
hi all, there was some discussion on this list about HP-UX and c++ modules. we had a mixture like this running with 2.0.43. we've recently upgraded to 2.0.44, and we're now experiencing problems at load time. i'm wondering if anyone is running 2.0.44 on HP-UX with c++ modules successfully, cheers, andy

Jeff Trawick | 1 Apr 2003 02:57

Re: Apache 2.0.45 -alpha tarball candidates for testing

William A. Rowe, Jr. wrote:
> The release tarball (and win32 .zip file) are now available for testing from;
> 
>   http://httpd.apache.org/dev/dist/

layouts (and thus the Apache binbuild mechanism) have some problems... 
can anybody get it to work?

FreeBSD 5.0:

unpack the alpha then
$ cd httpd-2.0.45
$ build/binbuild.sh
...
APR Version: 0.9.3
** Error: unable to find layout Apache
configure failed for srclib/apr
...

RH 6.2:

same error

Justin Erenkrantz | 1 Apr 2003 06:56
Gravatar

Re: Apache 2.0.45 -alpha tarball candidates for testing

--On Monday, March 31, 2003 7:57 PM -0500 Jeff Trawick <trawick <at> attglobal.net> 
wrote:

> layouts (and thus the Apache binbuild mechanism) have some problems... can
> anybody get it to work?
>
> FreeBSD 5.0:
>
> unpack the alpha then
> $ cd httpd-2.0.45
> $ build/binbuild.sh
> ...
> APR Version: 0.9.3
> ** Error: unable to find layout Apache
> configure failed for srclib/apr
> ...

The error message is correct.  APR doesn't have the layout called "Apache" - I 
don't think it ever has.  I think prior versions silently ignored this 
condition and just punted - which actually might have worked...

The workaround is to either use a layout that all of them share, or manually 
expand the layout (via the configure options).

The real fix is to drop the layout option before we do the subdir configure 
invocations in APR_SUBDIR_CONFIG (untested, but sort of based on how autoconf 
drops --no-create/--no-recursion):

ap_configure_args=
ap_sep=
for ac_configure_arg in $ac_configure_args
do
  case "$ac_configure_arg" in
    --enable-layout=*)
      continue ;;
  esac
  ap_configure_args="$ap_configure_args$ap_sep'$ac_configure_arg'"
  ap_sep=" "
done

ap_configure_args would be used instead of ac_configure_args in the subdir 
configure invocation.

And, the way that APR_SUBDIR_CONFIG is called within httpd's configure.in 
should allow the layout-expanded options to still be passed.  -- justin


Gmane