March 24, 2005
View 182fea0e4b

- the type of lump->op changed from int to enum _hdr_types-t - the value of HDR_OTHER_T changed to 0 - the value of HDR_VIA1_t changed to 1 (to avoid conflict with value 0 given as parameter to *_lump functions)

Jan Janak authored on 24/03/2005 14:47:26
View efc090fd07

- changed misleading log message (courtesy of Norman Brandinger)

Jan Janak authored on 24/03/2005 12:14:39
August 24, 2004
View 53c7e0f19a

- Spelling checked - READMEs updated

Jan Janak authored on 24/08/2004 08:45:09
February 23, 2004
View 41f45c80b4

- del_nonshm_lump(lump_list) added - removes from lump's tree all lumps that are not marked as LUMPFLAG_SHMEM.

Bogdan-Andrei Iancu authored on 23/02/2004 16:40:49
October 28, 2003
View 549dba7373

- added paranoic checks in {anchor,del}_lump for exceeding the msg. size

Andrei Pelinescu-Onciul authored on 28/10/2003 15:26:16
October 20, 2003
View 4c37101e0f

- added body_lumps (jan patches + some changes) - anchor_lump & del_lump compare the offset with msg->eoh and add the lump to add_rm or body_lumps, so make sure to parse_headers(HDR_EOH) if you lumps might modify the body. - changed all the {anchor,del}_lump calls

Andrei Pelinescu-Onciul authored on 20/10/2003 17:10:13
April 1, 2003
View 73eec4b209

- added contional lumps support functions: insert_cond_lump_after ( anchor, COND_TYPE, lump_type) insert_cond_lump_before

Andrei Pelinescu-Onciul authored on 01/04/2003 15:54:16
March 31, 2003
View f15bede127

- removed msg->repl_add_rm (msg->add_rm is used now both for replies & requests) - tcp code now always tries to add bind address info even for temporary "sending" sockets (it tries to find first match in the listening sockets list) - new lump function: insert_subst_lump_after, insert_subst_lump_before (you can add now lumps that will be substituted with the rcv/sending ip addr, port or proto: see SUBST_{SND,RCV}_{IP,PORT,PROTO})

Andrei Pelinescu-Onciul authored on 31/03/2003 19:51:10
February 7, 2003
View ab130758c9

- more tcp stuff and a lot of merging w/ latest cvs - content-length is appended automatically to messages that cross from udp to tcp - tcp2udp and udp2tcp now work under heavy stress (e.g.: throttle 200, 10 ser processes on dual cpu) - tcp performance still sucks, some things like disabling Nagle are still not in yet (for better debugging)

Andrei Pelinescu-Onciul authored on 07/02/2003 17:02:15
January 19, 2003
View 4d93f1459c

bunch of changes related to ability to process original request in reply_route: - textops: append_urihf(p,s) appends a hf with uri between p and s - core: supporting functions for duplicating a lump list from shmem to pkgmem (which allows further processing with pkgmem-oriented actions) - tm: reply_routes now creates a temporary, mostly pkg-mem-ed replica of shmem-ed request for additional processing

Jiri Kuthan authored on 19/01/2003 01:37:45
November 11, 2002
View e61f4ba3df

- openbsd (3.2) support [ changes in fastlock.h, ser_module.c, tm/t_hooks.c, Makefiles ] - texops replace/search_append bugs fixed

Andrei Pelinescu-Onciul authored on 11/11/2002 21:34:53
September 19, 2002
View 7dd0b34273

GPLization banner introduced to *.[hc] files

Jiri Kuthan authored on 19/09/2002 12:23:52
May 31, 2002
View 1400b77288

ser_error processing, ipv6-ization of TM, new TM callbacks; not stable yet (serial forking is screwed up)

Jiri Kuthan authored on 31/05/2002 01:59:06
March 2, 2002
View e22bbdb8e1

- fixed almost all warnings in ser core (gcc-3.0 -Wall, icc, sun cc). (=> fixed a lot of format string bugs) - replaced some of the remaining mallocs w/ pkg_malloc

Andrei Pelinescu-Onciul authored on 02/03/2002 02:20:00
January 7, 2002
View dda9dab1f8

performance and bug fixes

Jiri Kuthan authored on 07/01/2002 04:39:49