Summary: | sys-apps/gawk-4.1.1[readline] - build fails on Gentoo/FreeBSD | ||
---|---|---|---|
Product: | Gentoo/Alt | Reporter: | Yuta SATOH <nigoro.dev> |
Component: | FreeBSD | Assignee: | Gentoo/BSD Team <bsd+disabled> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | 4glitch, idarktemplar, kumba, prefix |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | FreeBSD | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 462580 | ||
Attachments: | bad patch for gawk-4.1.1/configure |
Description
Yuta SATOH
2014-04-12 05:27:01 UTC
(In reply to Yuta SATOH from comment #0) > Gentoo/FreeBSD does not have libtermcap.so. > However, gawk-4.1.1 requires it. > Reproducible: Always > > Steps to Reproduce: > 1. emerge sys-apps/gawk > 2. > 3. > > Actual Results: ... /usr/lib/gcc/i686-gentoo-freebsd9.1/4.6.3/../../../../i686-gentoo-freebsd9.1/bin/ld: cannot find -ltermcap collect2: error: ld returned 1 exit status ... I have the same result on g/fbsd-9.1. Created attachment 381574 [details, diff]
bad patch for gawk-4.1.1/configure
This is not good patch.
I used to make stages 1,2,3...
USE=-readline gets around this issue. I don't think that we would need readline support and recommend putting "sys-apps/gawk readline" in profiles/package.use.mask + 30 Jun 2015; Lars Wendler <polynomial-c@gentoo.org> gawk-4.1.3.ebuild, + +files/gawk-4.1.3-bsd_configure_readline.patch: + Fixed build with readline support on bsd (bug #507468). + Sorry this took so long. EPATCH_OPTS="-Z" does fail here, causing autotools to be run by Makefile.
patch's output (hidden by epatch) actually is:
> patching file configure
> Hunk #1 succeeded at 10825 (offset -7 lines).
> Not setting time of file configure (contents mismatch)
> patching file m4/readline.m4
> patch unexpectedly ends in middle of line
> Hunk #1 succeeded at 92 with fuzz 1.
> Not setting time of file m4/readline.m4 (contents mismatch)
Proposing:
diff -u -r1.2 gawk-4.1.3.ebuild
--- gawk-4.1.3.ebuild 30 Jun 2015 20:31:53 -0000 1.2
+++ gawk-4.1.3.ebuild 1 Jul 2015 10:44:08 -0000
@@ -32,7 +32,6 @@
-e '/check-recursive all-recursive: check-for-shared-lib-support/d' \
extension/Makefile.in || die
- EPATCH_OPTS="-Z" \
epatch "${FILESDIR}/${P}-bsd_configure_readline.patch" #507468
}
diff -u -r1.1 gawk-4.1.3-bsd_configure_readline.patch
--- files/gawk-4.1.3-bsd_configure_readline.patch 30 Jun 2015 20:31:53 -0000 1.1
+++ files/gawk-4.1.3-bsd_configure_readline.patch 1 Jul 2015 10:44:08 -0000
@@ -50,9 +50,9 @@
;;
esac
---- gawk-4.1.3/m4/readline.m4 2015-04-05 06:19:19.000000000 -0400
-+++ gawk-4.1.3/m4/readline.m4 2015-04-05 06:19:19.000000000 -0400
-@@ -92,7 +92,7 @@
+#--- gawk-4.1.3/m4/readline.m4 2015-04-05 06:19:19.000000000 -0400
+#+++ gawk-4.1.3/m4/readline.m4 2015-04-05 06:19:19.000000000 -0400
+#@@ -92,7 +92,7 @@
if test $_found_readline = yes ; then
case $host_os in
@@ -60,4 +60,4 @@
+ *bsd* ) AC_CHECK_LIB(termcap, tgetent, _combo="$_combo -ltermcap")
;;
esac
- AC_DEFINE(HAVE_LIBREADLINE,1,
\ No newline at end of file
+ AC_DEFINE(HAVE_LIBREADLINE,1,
(In reply to Michael Haubenwallner from comment #6) > EPATCH_OPTS="-Z" does fail here, causing autotools to be run by Makefile. > > patch's output (hidden by epatch) actually is: > > patching file configure > > Hunk #1 succeeded at 10825 (offset -7 lines). > > Not setting time of file configure (contents mismatch) > > patching file m4/readline.m4 > > patch unexpectedly ends in middle of line > > Hunk #1 succeeded at 92 with fuzz 1. > > Not setting time of file m4/readline.m4 (contents mismatch) > > Proposing: > > diff -u -r1.2 gawk-4.1.3.ebuild > --- gawk-4.1.3.ebuild 30 Jun 2015 20:31:53 -0000 1.2 > +++ gawk-4.1.3.ebuild 1 Jul 2015 10:44:08 -0000 > @@ -32,7 +32,6 @@ > -e '/check-recursive all-recursive: > check-for-shared-lib-support/d' \ > extension/Makefile.in || die > > - EPATCH_OPTS="-Z" \ > epatch "${FILESDIR}/${P}-bsd_configure_readline.patch" #507468 > } > > diff -u -r1.1 gawk-4.1.3-bsd_configure_readline.patch > --- files/gawk-4.1.3-bsd_configure_readline.patch 30 Jun 2015 20:31:53 > -0000 1.1 > +++ files/gawk-4.1.3-bsd_configure_readline.patch 1 Jul 2015 10:44:08 > -0000 > @@ -50,9 +50,9 @@ > ;; > esac > > ---- gawk-4.1.3/m4/readline.m4 2015-04-05 06:19:19.000000000 -0400 > -+++ gawk-4.1.3/m4/readline.m4 2015-04-05 06:19:19.000000000 -0400 > -@@ -92,7 +92,7 @@ > +#--- gawk-4.1.3/m4/readline.m4 2015-04-05 06:19:19.000000000 -0400 > +#+++ gawk-4.1.3/m4/readline.m4 2015-04-05 06:19:19.000000000 -0400 > +#@@ -92,7 +92,7 @@ > > if test $_found_readline = yes ; then > case $host_os in > @@ -60,4 +60,4 @@ > + *bsd* ) AC_CHECK_LIB(termcap, tgetent, _combo="$_combo > -ltermcap") > ;; > esac > - AC_DEFINE(HAVE_LIBREADLINE,1, > \ No newline at end of file > + AC_DEFINE(HAVE_LIBREADLINE,1, I can't seem to replicate the issue here. I've combinations of the patch and ebuild changes, and I have yet to see autotools interference. Though, Yutah SATOH (nigoro) has. nigoro, can you confirm that Michael's (haubi's) changes fix it on your end? + 02 Jul 2015; Lars Wendler <polynomial-c@gentoo.org> + files/gawk-4.1.3-bsd_configure_readline.patch: + Fixed gawk-4.1.3-bsd_configure_readline.patch so it applies correctly with -Z + option. + Sorry for the inconveniences. (In reply to Lars Wendler (Polynomial-C) from comment #8) Seems to work now, thanks! > Sorry for the inconveniences. This is what a buildbot setup is for ;) |