Hi Gavin!
Gavin Troy <[email protected]> wrote:
|Here's a segfault that I currently see when s-nail tries to \
Shit! Sorry!!
(IMAP, IMAP-cache, heh!
Does this mean i really need to keep that around all that time...)
|print out the last
|page of headers. When I move some mail out of my inbox it goes away, but it
|doesn't seem to be related to any particular mail. For the record, there are
|currently 870 messages there.
|I'm using v14.6.2-8-g68a9dc9 (and I haven't tested whether \
|stable is affected).
|
|#0 0x0000000000426097 in imap_getheaders (topp=<optimized \
|out>) at imap.c:1847
|#1 0x000000000040ecb7 in c_headers (v=<optimized out>) at cmd1.c:954
|#2 0x0000000000435af7 in announce (printheaders=<optimized \
|out>) at lex.c:1142
|#3 0x000000000043d899 in _rcv_mode (folder=<optimized out>, \
|Larg=<optimized out>) at main.c:539
|#4 main (argc=<optimized out>, argv=<optimized out>) at main.c:907
|
|(also attached NYD output)
Great! Thanks for that one! Much better than the sh...y
backtrace, so to say :)
I can reproduce it and i have pushed the fix (actually a partial
revert) to [master]; also attached.
Yes, the largest box i have tested over IMAP sofar is 507
messages, which is why i haven't seen the perfectly reproducable
test. Interestingly the code in question uses sections of 256
messages each. I'm looking into why the optimization doesn't
scale (so to say).
|Side issue: "make test" doesn't pass. Something related was \
|said in a previous
|list message, so maybe it's expected for now.
|
|> content:6: error: checksum mismatch (got 24505475 181)
NO!!! Really?? That is the same checksum that Gaetan reported!
What are you guys doing to this thing?? Strange.
Now i'm extremely curious what is going wrong there -- i already
spend almost two days on this, went insane, but for nothing!
Would you be so kind, patch in `gavin-test.diff' and send back the
`for-steffen.txt' that survives the test run? That would be very
much appreciated!
|In better news, I've not again seen the "memory corruption" issues I had
|before. So there's that. :)
<with-a-chirpy-voice/Yeeha/.
--steffen
Author: Steffen (Daode) Nurpmeso <[email protected]>
Date: 2014-03-12 12:59:07 +0100
Gavin Troy reported SEGV: partially revert [ac694bb]..
[ac694bb] (Two IMAP tries: fix SEGV and useless loads..,
2014-02-18) shouldn't make it into the release as such, because
the single line that i now revert was experimental (the `useless
loads' part).
I kept it in because a lot of testing showed no problems, and
i still need to inspect why boxes with more than a couple of
hundred mails cause the problem (all problem test boxes joined are
507 at the time of this writing, and cause no problems -- doubling
the box results in an immediate crash).
Thanks to Gavin Troy for reporting this including a NYD listing!
---
imap.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/imap.c b/imap.c
index f925230..66f820b 100644
--- a/imap.c
+++ b/imap.c
@@ -1844,7 +1844,7 @@ imap_getheaders(int volatile bot, int topp) /* TODO should take iterator!! */
for (i = bot; i <= topp; i += chunk) {
int j = i + chunk - 1;
- if (visible(message + j))
+ /* FIXME Gavin reported SEGV>800 msgs if (visible(message + j)) */
/*ok = */imap_fetchheaders(&mb, message, i, (j < topp ? j : topp));
if (interrupts)
onintr(0); /* XXX imaplock? */
diff --git a/cc-test.sh b/cc-test.sh
index 68924c4..3edb45e 100755
--- a/cc-test.sh
+++ b/cc-test.sh
@@ -423,6 +423,7 @@ SUB='Äbrä Kä?dä=brö Fü?di=bus? '\
MAILRC=/dev/null "${SNAIL}" -n#Sstealthmua \
-SPAGER="${cat}" -Spipe-text/plain="${cat}" > "${BODY}"
${sed} -e 1d < "${BODY}" > "${MBOX}"
+ cp "${MBOX} for-steffen.txt
cksum_test content:6 "${MBOX}" '3062395510 181'
${rm} -f "${BODY}" "${MBOX}"
------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and their
applications. Written by three acclaimed leaders in the field,
this first edition is now available. Download your free book today!
http://p.sf.net/sfu/13534_NeoTech
_______________________________________________
S-nail-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/s-nail-users
