[EMAIL PROTECTED] (Ken Williams) wrote:
>[EMAIL PROTECTED] (Doug MacEachern) wrote:
>>don't know why they're so different, but this looks like part of the
>>problem, since you do have unistd_h
>>i would try changing this part to:
>>
>>#ifndef HAVE_UNISTD_H
>>#define HAVE_UNISTD_H 1
>>#endif
>
>Yeah, that worked. I also changed the HAVE_SYS_PARAM_H def, which I
>assumed would also be needed.
>
>So I guess it's a mystery why those definitions were different in the
>two builds. Any hint for where to start looking?
I did some poking around - the ./configure command issued during the
mod_perl build process is:
CC="env LD_RUN_PATH=/System/Library/Perl/darwin/CORE cc" CFLAGS=" -g
-pipe -pipe -fno-common -DHAS_TELLDIR_PROTOTYPE -fno-strict-aliasing"
./configure --activate-module=src/modules/perl/libperl.a
--disable-rule=EXPAT --with-layout=Darwin
If I run that manually in apache's build directory, the same
definitional errors occur in ap_config_auto.h. However, if I omit the
CC="..." stuff:
CFLAGS=" -g -pipe -pipe -fno-common -DHAS_TELLDIR_PROTOTYPE
-fno-strict-aliasing" ./configure
--activate-module=src/modules/perl/libperl.a --disable-rule=EXPAT
--with-layout=Darwin
then I get the correct set of options in ap_config_auto.h, and the rest
of the build proceeds fine and all tests pass. I'm out of my league
when I try to figure out why that should make a difference.
Isn't this CC="..." unnecessary when building statically? That's what
the Config.pm POD seems to suggest.
------------------- -------------------
Ken Williams Last Bastion of Euclidity
[EMAIL PROTECTED] The Math Forum
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
