gcc-source: rewrite do_preconfigure() in shell

Message ID 1488471409-11114-1-git-send-email-ross.burton@intel.com
State New
Headers show

Commit Message

Ross Burton March 2, 2017, 4:16 p.m.
This function is essentially just four lines of shell, that for some reason was
written in Python.

Signed-off-by: Ross Burton <ross.burton@intel.com>

---
 meta/recipes-devtools/gcc/gcc-source.inc | 25 +++++++++++--------------
 1 file changed, 11 insertions(+), 14 deletions(-)

-- 
2.8.1

-- 
_______________________________________________
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Comments

Richard Purdie March 2, 2017, 4:21 p.m. | #1
On Thu, 2017-03-02 at 16:16 +0000, Ross Burton wrote:
> This function is essentially just four lines of shell, that for some

> reason was

> written in Python.

> 

> Signed-off-by: Ross Burton <ross.burton@intel.com>

> ---

> meta/recipes-devtools/gcc/gcc-source.inc | 25 +++++++++++-----------


I suspect this will break in the various signature checks and cause
rebuilds when things shouldn't.

http://git.yoctoproject.org/cgit.cgi/poky/commit/meta/?id=f07b7c0da1748afc35fbb2ead7ae10ec66c7c428

hints at why:

"It needs to be a python task to avoid lots of shell exported variables as dependencies."

which should of course be a comment in the code.

I wish we had a better way to control exported variables to shell tasks
which would be a better fix but right now we don't.

Cheers,

Richard
-- 
_______________________________________________
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Patch

diff --git a/meta/recipes-devtools/gcc/gcc-source.inc b/meta/recipes-devtools/gcc/gcc-source.inc
index fab99f2..7dcdb75 100644
--- a/meta/recipes-devtools/gcc/gcc-source.inc
+++ b/meta/recipes-devtools/gcc/gcc-source.inc
@@ -1,8 +1,8 @@ 
-deltask do_configure 
-deltask do_compile 
-deltask do_install 
+deltask do_configure
+deltask do_compile
+deltask do_install
 deltask do_populate_sysroot
-deltask do_populate_lic 
+deltask do_populate_lic
 RM_WORK_EXCLUDE += "${PN}"
 
 inherit nopackages
@@ -18,20 +18,17 @@  INHIBIT_DEFAULT_DEPS = "1"
 DEPENDS = ""
 PACKAGES = ""
 
-python do_preconfigure () {
-    import subprocess
-    cmd = d.expand('cd ${S} && PATH=${PATH} gnu-configize')
-    subprocess.check_output(cmd, stderr=subprocess.STDOUT, shell=True)
+do_preconfigure() {
+    (cd ${S} && gnu-configize) || die "failure in running gnu-configize"
+
     # See 0044-gengtypes.patch, we need to regenerate this file
-    bb.utils.remove(d.expand("${S}/gcc/gengtype-lex.c"))
-    cmd = d.expand("sed -i 's/BUILD_INFO=info/BUILD_INFO=/' ${S}/gcc/configure")
-    subprocess.check_output(cmd, stderr=subprocess.STDOUT, shell=True)
+    rm -f ${S}/gcc/gengtype-lex.c
+
+    sed -i 's/BUILD_INFO=info/BUILD_INFO=/' ${S}/gcc/configure
 
     # Easiest way to stop bad RPATHs getting into the library since we have a
     # broken libtool here (breaks cross-canadian and target at least)
-    cmd = d.expand("sed -i -e 's/hardcode_into_libs=yes/hardcode_into_libs=no/' ${S}/libcc1/configure")
-    subprocess.check_output(cmd, stderr=subprocess.STDOUT, shell=True)
+    sed -i -e 's/hardcode_into_libs=yes/hardcode_into_libs=no/' ${S}/libcc1/configure
 }
 addtask do_preconfigure after do_patch
 do_preconfigure[depends] += "gnu-config-native:do_populate_sysroot autoconf-native:do_populate_sysroot"
-