November 22, 2007
View 87c35a32ce

- updated Changelog and package info (ver. no.)

Andrei Pelinescu-Onciul authored on 22/11/2007 16:22:05
October 10, 2007
View c10674ded3

- the cpu is no longer hard-wired to strongarm1100 on arm arch. (strongarm is still the default, but it can be overwritten by setting CPU in the environment or make command line) - libfl dependency removed (makes cross compiling easier)

Andrei Pelinescu-Onciul authored on 10/10/2007 20:05:15
View 6c4eb49191

- backported fix: get_send_socket bug in some unlikely circumstances (a socket of a diff. proto is forced and find_si cannot fix it to a corresp. socket of the correct proto)

Andrei Pelinescu-Onciul authored on 10/10/2007 19:40:21
January 25, 2007
View 4b1bb28546

- variable timer fix: variable timers (avps) won't be exteneded anymore in some special situations. Tested with random generated timer values; minimal perfromance impact. Performance maniacs can turn the new code off modparam("tm", "var_timers", 0). Thanks to Bogdan Iancu <> for the excelent bug report.

Andrei Pelinescu-Onciul authored on 25/01/2007 16:14:49
December 15, 2006
View 787006ed8d

- dst_uri buffer overflow check, discovered in openser (the fix is slightly different due to slight changes in the append_branch function) Note: this is not such a big problem as it might sound, the only module using append_branch() with a non zero dst_uri is registrar and in this case it sets dst_uri to the received address which is always < 1024 (so it's always safe).

Andrei Pelinescu-Onciul authored on 15/12/2006 09:12:38
December 4, 2006
View f095ab5d41

- gcc 4.x minline-all-stringops removed from the arm & alpha architectures (not supported)

Andrei Pelinescu-Onciul authored on 04/12/2006 21:40:08
November 29, 2006
View 96be40ecc3

- reverted last commit tm changes, which got in there by mistake (it was the var timer fix, which will be commited later when I'm confident enought that it will not break anything)

Andrei Pelinescu-Onciul authored on 29/11/2006 14:41:39
View 5fbf103fd7

- fix for free_rdata_list() which used to access the "next" pointer after freeing the current elements (ported from Ottendorf) Credits & patch: Jan Andres <>

Andrei Pelinescu-Onciul authored on 29/11/2006 14:39:05
November 24, 2006
View 1101923f1f

- usrloc hash fix ported from ser 0.10.99-dev60 (Ottendorf)

Andrei Pelinescu-Onciul authored on 24/11/2006 17:38:44
November 22, 2006
View c15157fb56

- fastlock arm fix: arm early clobber added: according to the swp instruction specification the address register must be != from the other 2 . Fix from openser, originally sent by Julien Blache <>)$

Andrei Pelinescu-Onciul authored on 22/11/2006 18:46:53
October 25, 2006
View bb8e2a3579

- print warnings number if any

Andrei Pelinescu-Onciul authored on 25/10/2006 19:33:32
August 28, 2006
View 73194f939d

- deadlock when t_relay-ing a message from the failure_route fixed (e2e_cancel uses now t_reply_unsafe when called from the failure_route)

Andrei Pelinescu-Onciul authored on 28/08/2006 11:21:02
August 25, 2006
View 0ffeabb831

- fixed set_advertised_port

Andrei Pelinescu-Onciul authored on 25/08/2006 14:02:33
July 13, 2006
View ccba9ce6ca

- fixed possible memory leak when receiving bad responses for srv or naptr requests (it can happen if more records are returned and the lenght of these records is tweaked to exceed tha packet size)

Andrei Pelinescu-Onciul authored on 13/07/2006 18:35:32
June 13, 2006
View ef54ef0ac4

'fixed the fix' after andrei's directions.

Raphael Coeffic authored on 13/06/2006 13:31:48