I can confirm this fixes the leak on my end. Thanks!
> I'm slightly worried that ASAN tests will prove to be flaky.
If I recall correctly, the Chromium project runs ASAN in their CI
pipelines and it's pretty stable. Hopefully that's some evidence towards
their stability.
Austin
--
Hi,
https://nmbug.notmuchmail.org/nmweb/static/js/jquery-ui.js returns a 404
so nmweb's date pickers don't work. Based on the symlink in the repo, it
looks like a deployment issue.
Hope this helps!
Austin
--
https://austinray.io
Open Source Maintainer, Software Engineer, Keyboard Enthusiast
Hi,
libnotmuch 5.5.0 (notmuch 0.34.2) leaks memory when opening a database
with 'notmuch_database_open_with_config()' and cleaning it up with
'notmuch_database_destroy()'. I've included a reproducer program (based
on an existing notmuch test) and ASAN output below.
Replacing
-by: Austin Ray
---
contrib/notmuch-web/nmweb.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/contrib/notmuch-web/nmweb.py b/contrib/notmuch-web/nmweb.py
index 9e337436..0754a502 100755
--- a/contrib/notmuch-web/nmweb.py
+++ b/contrib/notmuch-web/nmweb.py
@@ -14,7 +14,7 @@ from
Hi David,
This resolves both the non-NULL pointer and memory leak reported by
ASAN.
Thanks!
Austin
--
https://austinray.io
Open Source Maintainer, Software Engineer, Keyboard Enthusiast
GPG: 0127 ED83 B939 CCC9 8082 476E 1AA0 B115 C8AC 2C9E
signature.asc
Description: PGP signature
Following up on this, ASAN reports a memory leak on a failure. Using
'notmuch_database_close()' or 'notmuch_database_destroy()' results in a
SEGFAULT.
I've updated my reproducer program to free the database on non-failure
cases so it's easier to see the memory leak.
#include
#include
int
Submitting this bug report per bremner's request (thanks for the assist
today!)
The documentation for 'notmuch_database_open_with_config()' states:
> In case of any failure, this function returns an error status and
> sets *database to NULL.
However, it's possible to trigger a failure and leave
> I guess "-v 2" makes sense here. It's just a hint that the new patch
> obsoletes a previous one.
Thanks! I appreciate the help!
Austin
signature.asc
Description: PGP signature
___
notmuch mailing list -- notmuch@notmuchmail.org
To unsubscribe send
Both notmuch_database_open() and notmuch_database_open_verbose()'s
documentation state they call notmuch_database_open_with_config() with
config_path=NULL; however, their implementations pass an empty string.
The empty string is the correct value to maintain their original
behavior of not loading
> So maybe it's a doc bug?
That sounds good to me.
Since notmuch_database_open() has a similar doc bug, would you like the
corrections in one or two patches during the resubmit?
Also, since I'm new to this git workflow, should I use the "-vN" option
for the resubmit or leave it off as this
Notmuch 0.32 corresponds to libnotmuch 5.4 as indicated by docstrings;
however, the minor number wasn't bumped. Any libnotmuch downstream
consumer using the LIBNOTMUCH_CHECK_VERSION macro to support multiple
versions won't be able to access the new 5.4 functions.
Signed-off-by: Austin Ray
11 matches
Mail list logo