On 06/14/2016 09:03 AM, Martin Jansa wrote:
* otherwise there is a lot of warnings about missing close on file descriptor
Signed-off-by: Martin Jansa <[email protected]>
---
meta/classes/icecc.bbclass | 14 ++++++++++----
1 file changed, 10 insertions(+), 4 deletions(-)
diff --git a/meta/classes/icecc.bbclass b/meta/classes/icecc.bbclass
index e1c06c4..cb44007 100644
--- a/meta/classes/icecc.bbclass
+++ b/meta/classes/icecc.bbclass
@@ -47,7 +47,8 @@ def get_cross_kernel_cc(bb,d):
# evaluate the expression by the shell if necessary
if '`' in kernel_cc or '$(' in kernel_cc:
- kernel_cc = os.popen("echo %s" % kernel_cc).read()[:-1]
+ import subprocess
There are two imports like this, would it make sense to place just once
on top?
Also, now that we switch to subprocess.check_output, I wonder if we need
to convert it to string (the output of this method) before proceeding.
+ kernel_cc = subprocess.check_output("echo %s" % kernel_cc, shell=True)
kernel_cc = d.expand(kernel_cc)
kernel_cc = kernel_cc.replace('ccache', '').strip()
@@ -220,9 +221,14 @@ def icecc_get_and_check_tool(bb, d, tool):
# PATH or icecc-create-env script will silently create an invalid
# compiler environment package.
t = icecc_get_tool(bb, d, tool)
- if t and os.popen("readlink -f %s" % t).read()[:-1] == get_icecc(d):
- bb.error("%s is a symlink to %s in PATH and this prevents icecc from
working" % (t, get_icecc(d)))
- return ""
+ if t:
+ import subprocess
+ link_path = subprocess.check_output("readlink -f %s" % t, shell=True)
+ if link_path == get_icecc(d):
+ bb.error("%s is a symlink to %s in PATH and this prevents icecc from
working" % (t, get_icecc(d)))
+ return ""
+ else:
+ return t
else:
return t
--
_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.openembedded.org/mailman/listinfo/openembedded-core
