aboutsummaryrefslogtreecommitdiffstats
path: root/bitbake
diff options
context:
space:
mode:
authorRichard Purdie <richard.purdie@linuxfoundation.org>2012-05-24 14:23:44 +0100
committerRichard Purdie <richard.purdie@linuxfoundation.org>2012-05-25 11:15:12 +0100
commite6d9af19d6ef94522e5c53eb84c490e7b5473c69 (patch)
tree3fa0c97270cd762551087cfd3731ca91475f5fda /bitbake
parent0b14db452413df414fbadf14d6dae287bc5ae9ca (diff)
downloadopenembedded-core-contrib-e6d9af19d6ef94522e5c53eb84c490e7b5473c69.tar.gz
bitbake/exceptions: Handle reports from the field of exception code failures
Despite using python 2.6, there have been reports of issues where bitbake is printing tracebacks with errors in the exception handling code. This was masking the real error. Since we need to do whatever we can to give the user good feedback about errors, detect the tuple instead of namedtuple case and don't fault in the exception handler but just give up trying to traceback any further. In the reported cases, this gives a message the user can then understand. (Bitbake rev: 9ec0429271e68527a55fc123dea5a1b959c6ec3b) Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Diffstat (limited to 'bitbake')
-rw-r--r--bitbake/lib/bb/exceptions.py9
1 files changed, 8 insertions, 1 deletions
diff --git a/bitbake/lib/bb/exceptions.py b/bitbake/lib/bb/exceptions.py
index 4dd3e2cc31..f182c8fd62 100644
--- a/bitbake/lib/bb/exceptions.py
+++ b/bitbake/lib/bb/exceptions.py
@@ -32,7 +32,14 @@ class TracebackEntry(namedtuple.abc):
def _get_frame_args(frame):
"""Get the formatted arguments and class (if available) for a frame"""
arginfo = inspect.getargvalues(frame)
- if not arginfo.args:
+
+ try:
+ if not arginfo.args:
+ return '', None
+ # There have been reports from the field of python 2.6 which doesn't
+ # return a namedtuple here but simply a tuple so fallback gracefully if
+ # args isn't present.
+ except AttributeError:
return '', None
firstarg = arginfo.args[0]