Bug 377279 – app-emulation/libvirt-9999 must run bootstrap

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Michal Privoznik 2011-08-01 13:16:10 UTC]

Libvirt requires bootstrapping script to be run to create configure script. Moreover, this script clones subproject repo in the right place.

Reproducible: Always

Steps to Reproduce:
1.make sure you don't have libvirt git cloned in distfiles
2.emerge =libvirt-9999
3.observe error
Actual Results:  
emerge fail

Expected Results:  
emerge will continue and finish successfuly

Fix:

--- /usr/portage/app-emulation/libvirt/libvirt-9999.ebuild      2011-07-20 17:01:03.000000000 +0200
+++ /usr/local/portage/app-emulation/libvirt/libvirt-9999.ebuild        2011-08-01 14:29:52.840004834 +0200
@@ -99,6 +99,15 @@ src_prepare() {
                        "${S}"/tests/commanddata/test8.log || die
        fi
 
+       ./bootstrap || die "bootstrap failed"
+       (
+                       git submodule status | sed 's/^[ +-]//;s/ .*//'
+                       git hash-object bootstrap.conf
+       ) >.git-module-status
+       # this has already been done by bootstrap but we're required to redo it to
+       # remove maintainer mode
+       eautoreconf
+
        [[ -n ${AUTOTOOLIZE} ]] && eautoreconf
 }
 
@@ -190,6 +199,9 @@ src_configure() {
                --with-remote \
                --localstatedir=/var \
                --with-remote-pid-file=/var/run/libvirtd.pid
+
+       # HACK, restore gnulib's config.sub and config.guess
+       (cd .gnulib && git reset --hard >/dev/null)
 }
 
 src_test() {

[Comment 1 Aaron Rosen 2012-02-17 21:00:45 UTC]

I ran into this same issue. I changed the ebuild using Michal's changes below. Here is an updated diff since the ebuild has changed since his post. 

diff --git a/libvirt-9999.ebuild b/usr/local/portage/app-emulation/libvirt/libvirt-9999.ebuild
index ea3f038..e95a6a3 100644
--- a/libvirt-9999.ebuild
+++ b/usr/local/portage/app-emulation/libvirt/libvirt-9999.ebuild
@@ -96,6 +96,17 @@ pkg_setup() {
 }
 
 src_prepare() {
+
+       ./bootstrap || die "bootstrap failed"
+       (
+               git submodule status | sed 's/^[ +-]//;s/ .*//'
+               git hash-object bootstrap.conf
+       ) >.git-module-status
+       # this has already been done by bootstrap but we're required to redo
+
+       #remove maintainer mode
+       eautoreconf
+
        [[ -n ${BACKPORTS} ]] && \
                EPATCH_FORCE=yes EPATCH_SUFFIX="patch" EPATCH_SOURCE="${S}/patches" \
                        epatch
@@ -193,6 +204,11 @@ src_configure() {
                --docdir=/usr/share/doc/${PF} \
                --with-remote \
                --localstatedir=/var
+
+
+       # HACK, restore gnulib's config.sub and config.guess
+       (cd .gnulib && git reset --hard >/dev/null)
+
 }
 
 src_test() {

[Comment 2 Julian Ospald 2012-02-17 21:08:54 UTC]

Created attachment 302295 [details, diff]
bootstrap.patch

[Comment 3 Aaron Rosen 2012-02-17 21:12:01 UTC]

Yup, hasufell patch is better, the bootstrap should be run after .
  
  [[ -n ${BACKPORTS} ]] && \
        EPATCH_FORCE=yes EPATCH_SUFFIX="patch" EPATCH_SOURCE="${S}/patches" \
            epatch

[Comment 4 Julian Ospald 2012-02-17 21:40:45 UTC]

Created attachment 302301 [details, diff]
bootstrap.patch

[Comment 5 Doug Goldstein (RETIRED) 2012-03-05 23:44:37 UTC]

Thanks all. Fixed in the tree.