[Citadel Development] (no subject)

[Freakdog]

 

Mon Feb 01 2010 08:21:37 AM EST from   dothebart @ Uncensored 

http://googleenterprise.blogspot.com/2010/01/modern-browsers-for-modern-applications.html
pull the plug!


I can understand IE 6, and even FF 2...but FF 3.0.x is still supported by Mozilla, at this point.

[Citadel Development] Citadel commit log: revision 8270

[dothebart]

r8270 | dothebart | 2010-02-01 14:22:33 -0500 (Mon, 01 Feb 2010) | 2 lines
Changed paths:
   M /trunk/webcit/serv_func.c

* don't free the buffer in here, its _not_ ours.
* abort if the server has gone away

[Citadel Development] Citadel commit log: revision 8269

[dothebart]

r8269 | dothebart | 2010-02-01 14:19:34 -0500 (Mon, 01 Feb 2010) | 2 lines
Changed paths:
   M /trunk/citadel/file_ops.c

* partialy revert r8246; we're working in units of one here, so the number of 
units read actualy _is_ the number of bytes read here.

[Citadel Development] (no subject)

[dothebart]

 

Mo Feb 01 2010 11:09:33 EST von   dothebart @ Uncensored 

 

Mo Feb 01 2010 10:22:40 EST von   dothebart @ Uncensored 

hm, I've seem to have broken the attachment downloading?


whew. this is strange.
I can trace citserver doing...
CC->download_fp = tmpfile(); if (CC->download_fp == NULL) return;  rv = fwrite(content, length, 1, CC->download_fp); fflush(CC->download_fp); rewind(CC->download_fp);  OpenCmdResult(filename, cbtype);
and... in file_ops cmd_read()
fseek(CC->download_fp, start_pos, 0); if (fread(buf, 1, bytes, CC->download_fp) == 1) {will go to
cprintf("%d %s\n", ERROR, strerror(errno));and print "success"
WTF?


Ok, this is a bug in the fread() bug correction.
fread will return 4096 here, not 1
most probably because of unitsize and units is swapped here in opposite to the other cases:
r8246 | ajc | 2010-01-18 00:12:55 +0100 (Mo, 18. Jan 2010) | 1 ZeileGeänderte Pfade: M /trunk/citadel/config.c M /trunk/citadel/file_ops.c* Fixed another fread() bug
so, should we revert the file_ops part of r8246?

[Citadel Development] (no subject)

[dothebart]

 

Mo Feb 01 2010 10:22:40 EST von   dothebart @ Uncensored 

hm, I've seem to have broken the attachment downloading?


whew. this is strange.
I can trace citserver doing...
CC->download_fp = tmpfile(); if (CC->download_fp == NULL) return;  rv = fwrite(content, length, 1, CC->download_fp); fflush(CC->download_fp); rewind(CC->download_fp);  OpenCmdResult(filename, cbtype);
and... in file_ops cmd_read()
fseek(CC->download_fp, start_pos, 0); if (fread(buf, 1, bytes, CC->download_fp) == 1) {will go to
cprintf("%d %s\n", ERROR, strerror(errno));and print "success"
WTF?

[Citadel Development] (no subject)

[dothebart]

hm, I've seem to have broken the attachment downloading?
in other news, a memory leak here:
==31277== 31 bytes in 1 blocks are definitely lost in loss record 429 of 529==31277==    at 0x4024C4C: malloc (vg_replace_malloc.c:195)==31277==    by 0x450E72F: strdup (strdup.c:43)==31277==    by 0x806FDA0: convert_field (internet_addressing.c:647)==31277==    by 0x806FECD: convert_internet_message (internet_addressing.c:756)==31277==    by 0x80A9B78: smtp_data (serv_smtp.c:723)==31277==    by 0x80AAF1C: smtp_command_loop (serv_smtp.c:893)==31277==    by 0x8056F17: worker_thread (sysdep.c:1020)==31277==    by 0x807401E: ctdl_internal_thread_func (threadsc:839)==31277==    by 0x4232584: start_thread (pthread_create.c:300)==31277==    by 0x456A2BD: clone (clone.S:130)
 
else if (!strcasecmp(key, "In-reply-to")) { if (msg->cm_fields['W'] == NULL) {        /* References: supersedes In-reply-to: */ msg->cm_fields['W'] = strdup(value); } processed = 1; }
which is 'W'
IG, since you've added that...  do you know a place where it could get overwritten without free?
 

[Citadel Development] (no subject)

[IGnatius T Foobar]

It's about time.  So we do have to figure out what code we have in our framework
that exists solely to support obsolete browsers.  That'll be the challenge. 
(Matt?  Thierry?  Anyone know for sure?)