[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]
[PATCH] Make the save-to-bugzilla dupe detection smarter.
- From: Chris Lumens <clumens redhat com>
- To: anaconda-devel-list redhat com
- Cc:
- Subject: [PATCH] Make the save-to-bugzilla dupe detection smarter.
- Date: Tue, 26 May 2009 15:01:55 -0400
This fixes two problems with how we determine tracebacks are the same based
on filename comparisons:
(1) Comparisons between python files on i386 and x86-64 tracebacks fail because
/usr/lib and /usr/lib64 are different, even though the files themselves are the
same.
(2) Comparisons between python files on the running system and in /tmp/updates
fail obviously. This is a little less concerning to fail on, but since the
lines of code are likely to change in an updates.img, duplicates should still
work properly.
---
exception.py | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/exception.py b/exception.py
index 3bb8781..7b09646 100644
--- a/exception.py
+++ b/exception.py
@@ -243,7 +243,7 @@ class AnacondaExceptionDump:
for (file, lineno, func, text) in [f[1:5] for f in self.stack]:
if type(text) == type([]):
text = "".join(text)
- s += "%s %s %s\n" % (file, func, text)
+ s += "%s %s %s\n" % (os.path.basename(file), func, text)
return hashlib.sha256(s).hexdigest()
--
1.6.1.3
[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]