[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gentoo-perl
Subject:    Re: [gentoo-perl] www-server/mod_perl-2.0.3 ebuild
From:       Davide Angelocola <davide.angelocola () tiscali ! it>
Date:       2007-02-06 23:03:37
Message-ID: 200702070003.38219.davide.angelocola () tiscali ! it
[Download RAW message or body]

On Tuesday 06 February 2007 20:28:55 Michael Cummings wrote:
> Davide Angelocola wrote:
> <snip>
>
> > That is ok. However I think that "die" should be better than "exit"
> > here in order to avoid the error "Can't open ...".
>
> Except wouldn't exit just exit the sub...? (bash is so much fun) Did you
> test this locally (probably not, I know :)?
I've also a patch (attached) that produces the following output:
# emerge mod_perl

>>> Emerging (1 of 1) www-apache/mod_perl-2.0.3 to /
 * mod_perl-2.0.3.tar.gz MD5 ;-) ...                                      [ 
ok ]
 * mod_perl-2.0.3.tar.gz RMD160 ;-) ...                                   [ 
ok ]
 * mod_perl-2.0.3.tar.gz SHA1 ;-) ...                                     [ 
ok ]
 * mod_perl-2.0.3.tar.gz SHA256 ;-) ...                                   [ 
ok ]
 * mod_perl-2.0.3.tar.gz size ;-) ...                                     [ 
ok ]
 * checking ebuild checksums ;-) ...                                      [ 
ok ]
 * checking auxfile checksums ;-) ...                                     [ 
ok ]
 * checking miscfile checksums ;-) ...                                    [ 
ok ]
 * checking mod_perl-2.0.3.tar.gz ;-) ...                                 [ 
ok ]
>>> Unpacking source...
 * You cannot build mod_perl on a threaded apache
 * with an unthreaded perl. You must either emerge
 * perl with ithreads in your USE flags, or emerge
 * apache without threading support

!!! ERROR: www-apache/mod_perl-2.0.3 failed.
Call stack:
  ebuild.sh, line 1613:   Called dyn_unpack
  ebuild.sh, line 750:   Called qa_call 'src_unpack'
  environment, line 3600:   Called src_unpack
  mod_perl-2.0.3.ebuild, line 38:   Called die

!!! (no error message)
!!! If you need support, post the topmost build error, and the call stack if 
relevant.
!!! A complete build log is located 
at '/var/tmp/portage/www-apache/mod_perl-2.0.3/temp/build.log'.

!!! This ebuild is from an overlay: '/usr/local/portage/dfa-overlay'

PS
the subject is wrong: it should be www-apache/mod-perl. Sorry :-)

-- 
Best Regards,
-- Davide Angelocola

["die.diff" (text/x-diff)]

--- mod_perl-2.0.3.ebuild.orig	2007-02-06 23:59:05.600764709 +0100
+++ mod_perl-2.0.3.ebuild	2007-02-06 23:59:45.670634199 +0100
@@ -31,11 +31,11 @@
 	ITHREADS="${useithreads}"
 	if [ "${INSTALLED_MPM}" != "prefork" ]; then
 		if [ "${ITHREADS}" == "undef" ]; then
-			ewarn "You cannot build mod_perl on a threaded apache"
-			ewarn "with an unthreaded perl. You must either emerge"
-			ewarn "perl with ithreads in your USE flags, or emerge"
-			ewarn "apache without threading support"
-			exit
+			eerror "You cannot build mod_perl on a threaded apache"
+			eerror "with an unthreaded perl. You must either emerge"
+			eerror "perl with ithreads in your USE flags, or emerge"
+			eerror "apache without threading support"
+			die
 		fi
 	fi
 
@@ -45,14 +45,14 @@
 
 	# I am not entirely happy with this solution, but here's what's
 	# going on here if someone wants to take a stab at another
-	# approach.  When userpriv compilation is off, then the make
+	# approach.	 When userpriv compilation is off, then the make
 	# process drops to user "nobody" to run the test servers.  This
 	# server is closed, and then the socket is rebound using
-	# SO_REUSEADDR.  If the same user does this, there is no problem,
+	# SO_REUSEADDR.	 If the same user does this, there is no problem,
 	# and the socket may be rebound immediately.  If a different user
 	# (yes, in my testing, even root) attempts to rebind, it fails.
 	# Since the "is the socket available yet" code and the
-	# second-batch bind call both run as root, this will fail.  
+	# second-batch bind call both run as root, this will fail.
 
 	# The upstream settings on my test machine cause the second batch
 	# of tests to fail, believing the socket to still be in use.  I
@@ -85,16 +85,16 @@
 		MP_TRACE=1 \
 		MP_DEBUG=1 \
 		MP_USE_DSO=1 \
-		MP_APXS=/usr/sbin/apxs2  \
+		MP_APXS=/usr/sbin/apxs2	 \
 		INSTALLDIRS=vendor </dev/null || die
 
 	# reported that parallel make is broken in bug 30257
 	emake -j1 || die
 
-	# mcummings - disabling tests right, there's a bug i haven't found an override for
-	# rendhalver - i seem to have fixed this so turning them back on again. 
-	# yet that sets /tmp as your tmp dir
-	#hasq maketest $FEATURES && src_test
+	# mcummings - disabling tests right, there's a bug i haven't found
+	# an override for rendhalver - i seem to have fixed this so
+	# turning them back on again. yet that sets /tmp as your tmp dir
+	# hasq maketest $FEATURES && src_test
 }
 
 src_test() {

-- 
gentoo-perl@gentoo.org mailing list


[prev in list] [next in list] [prev in thread] [next in thread]