After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 202529 - imap commands can be too long
imap commands can be too long
Status: RESOLVED DUPLICATE of bug 216927
Product: evolution
Classification: Applications
Component: Mailer
pre-1.5 (obsolete)
Other other
: Normal major
: ---
Assigned To: Jeffrey Stedfast
Evolution QA team
Depends on:
Blocks:
 
 
Reported: 2001-05-01 14:24 UTC by Vladimir Vukicevic
Modified: 2002-01-15 18:47 UTC
See Also:
GNOME target: ---
GNOME version: ---



Description Vladimir Vukicevic 2001-05-01 14:24:21 UTC
Select a good amount of messages, say a few thousand.  Hit Delete.

Error while performing opration:
Unexpected response from IMAP server: A00295 BAD Command line too long
Comment 1 Dan Winship 2001-05-06 16:20:14 UTC
Oh, there was a bug that I've fixed that was making it say
1,2,3,4,5,6,7,... instead of 1:1000 when setting flags, so that's
probably what caused the bug in the case you mentioned, although
you could still trigger it by deleting a large number of
non-contiguous messages.
Comment 2 Luis Villa 2001-06-25 16:59:21 UTC
vlad: can you still duplicate this? I'm going to mark FIXED as per
Dan's comment, but we can re-open easily.
Comment 3 Luis Villa 2001-10-30 17:15:31 UTC
yeah, so the problem here was definitely non-contigous messages (in
this case, only bugzilla bug changes were moved, while new bugs and
'real' mail were left untouched.) I've put in nfs the trace of the
error file. I'm also upping the priority a touch, since it makes the
program unusable and doesn't deal with the error message generated
well at all. :/
Comment 4 Luis Villa 2001-10-30 17:16:12 UTC
doh. nfs location is /nfs/louie/bug_2529_trace.txt .
Comment 5 Luis Villa 2001-11-26 17:01:40 UTC
Because of the decision to remap 1.1->1.2 and 1.2->1.4, I'm going to be
moving a large number of bugs around in the bugzilla. You can just
search on 'body contains' 'Because of the decision to remap' and mark
all as read. Please direct all questions about this change to
evolution@ximian.com, not the bug.
Luis
Comment 6 Jeffrey Stedfast 2002-01-11 19:31:49 UTC
From Courier-imapd sources:

#define IT_MAX_ATOM_SIZE 16384

(IT == short for IMAP_TOKEN it seems)

not sure what uw.imapd uses as it's maximum token size...

this is just plain sad.
Comment 7 Jeffrey Stedfast 2002-01-14 22:01:59 UTC
reassigning to me for 1.2

I've actually fixed most of this, there is only 1 location that still
needs fixing afaik which is imap_update_summary() in
camel-imap-folder.c

Well, "fixed" in so far as it will limit uid-set tokens to a
pre-defined size. As far as I know, no other tokens can get large
enough to matter. I guess we may still have to limit the entire
command-line limit to ~8k, but afaik - this shouldn't be an issue if
we limit the uid-set size to something reasonable. rfc2683 suggests 1k
but I think we can actually get away with something larger, like maybe
4-6k.

Currently I've defined UID_SET_LIMIT to -1 (aka unlimited) but we can
either hard-code this value to whatever we want or maybe special-case
the value based on the server impl? >:-)

For Courier, we know that a token must be <16k but that the
command-line is not limited.

For UW, we know that the token size doesn't matter so long as the
entire command-line is <8k

For an unknown server, we could limit each token to say 1k so that we
are sure to work...

and if we ever write Camel-imapd, we can limit it to some other size
(or unlimited if we write camel-imapd to not impose any limitations)
Comment 8 Jeffrey Stedfast 2002-01-15 18:45:33 UTC
I've just fixed imap_update_summary to respect UID_SET_LIMIT now. So
this is the last of the functions that needed updating.

The only part of this bug that can possibly still go wrong is maybe a
command-line could be too long despite the limitation of the uid-set
token. The only place I can think of that could possibly cause a
command-line to be too long now is the SEARCH command for like BODY
and HEADER searches (since they take user input). But this can't
possibly be fixed until we use Zucchi's ImapEngine so I am now marking
this as a duplicate of "imap rewrite"
Comment 9 Jeffrey Stedfast 2002-01-15 18:47:45 UTC

*** This bug has been marked as a duplicate of 216927 ***