Callweaver

Fixed unsecure code. When dumping the object, use '%s' and pass the string,...

Wed, 05 Jan 2011 12:50:30 +0000

Commit by karvan on rel :: r5717 callweaver/1.2/include/callweaver/cwobj.h: (link)

Fixed unsecure code.

When dumping the object, use '%s' and pass the string, don't use it as a format element.

Remove an extraneous backslash

Thu, 15 Jul 2010 15:58:18 +0000

Commit by mjagdis :: r5716 callweaver/corelib/callweaver_expr2.l: (link)

Remove an extraneous backslash

Update acmacros/ax_check_openssl.m4

Mon, 07 Jun 2010 16:16:15 +0000

Commit by mjagdis :: r5715 callweaver/acmacros/ax_check_openssl.m4: (link)

Update acmacros/ax_check_openssl.m4

pbx_ael should not abuse cw_expr() as a syntax checker Expression evaluation is...

Sat, 05 Jun 2010 10:57:50 +0000

Commit by mjagdis :: r5714 callweaver/ (3 files in 2 dirs): (link)

pbx_ael should not abuse cw_expr() as a syntax checker

Expression evaluation is a two stage process, substitution then
evaluation. Just calling cw_expr() is simply wrong because it isn't
parsing what it would under true evaluation. Worse, you don't want to
expand or evaluate things with possible side effects when simply loading
a dialplan!

Removing this means that some syntax errors will only be detectable at
run-time rather than (possibly) load-time. But that's the nature of the
beast. Dialplan isn't a language. AEL is a text transformation tool that
transforms text into dialplan. AEL is not a language either. IMHO no one
should use AEL and it shouldn't even exist. At the very least it should
exist as a separate "compiler" and not be embedded.

When converting an arg to a number remember it may have an exponent

Fri, 04 Jun 2010 21:11:27 +0000

Commit by mjagdis :: r5713 callweaver/funcs/func_core.c: (link)

When converting an arg to a number remember it may have an exponent

Initialize the randomlock after the mutexattr has been initialized

Fri, 04 Jun 2010 21:11:19 +0000

Commit by mjagdis :: r5712 callweaver/corelib/utils.c: (link)

Initialize the randomlock *after* the mutexattr has been initialized

Use POSIX waitpid() rather than BSD wait4()

Fri, 04 Jun 2010 21:11:11 +0000

Commit by mjagdis :: r5711 callweaver/ (2 files in 2 dirs): (link)

Use POSIX waitpid() rather than BSD wait4()

"help" with no args should list everything, not crash callweaver

Fri, 04 Jun 2010 21:11:03 +0000

Commit by mjagdis :: r5710 callweaver/corelib/cli.c: (link)

"help" with no args should list everything, not crash callweaver

Removed check_expr It hasn't been built for ages and if anyone thinks what it...

Fri, 04 Jun 2010 21:10:55 +0000

Commit by mjagdis :: r5709 callweaver/utils/ (Makefile.am check_expr.c): (link)

Removed check_expr

It hasn't been built for ages and if anyone thinks what it does it
useful they should perhaps read the code and compare it with the reality
of the core substitution, expression evaluation and argument splitting
routines.

Add support for tracing object lifecycles

Fri, 04 Jun 2010 21:10:47 +0000

Commit by mjagdis :: r5708 callweaver/include/callweaver/object.h: (link)

Add support for tracing object lifecycles

Use a rwlock to protect the dialplan data structures rather than a simple mutex

Fri, 04 Jun 2010 21:10:38 +0000

Commit by mjagdis :: r5707 callweaver/corelib/ (callweaver.c pbx.c): (link)

Use a rwlock to protect the dialplan data structures rather than a simple mutex

Remove unused cruft from config handling

Fri, 04 Jun 2010 21:10:30 +0000

Commit by mjagdis :: r5706 callweaver/ (2 files in 2 dirs): (link)

Remove unused cruft from config handling

It doesn't matter what the byte is that we write to the alert pipe

Fri, 04 Jun 2010 21:10:22 +0000

Commit by mjagdis :: r5705 callweaver/corelib/channel.c: (link)

It doesn't matter what the byte is that we write to the alert pipe

Avoid accessing beyond the end of the iov list in cw_writev_all

Fri, 04 Jun 2010 21:10:14 +0000

Commit by mjagdis :: r5704 callweaver/corelib/utils.c: (link)

Avoid accessing beyond the end of the iov list in cw_writev_all

Fix CPU binding/speed setting when doing translation timings

Fri, 04 Jun 2010 21:10:06 +0000

Commit by mjagdis :: r5703 callweaver/corelib/translate.c: (link)

Fix CPU binding/speed setting when doing translation timings

Start up of the DUNDI network listener was broken I guess no one uses DUNDI...

Fri, 04 Jun 2010 21:09:58 +0000

Commit by mjagdis :: r5702 callweaver/pbx/pbx_dundi.c: (link)

Start up of the DUNDI network listener was broken

I guess no one uses DUNDI then?

Make sure we initialize the whole of a sockaddr_in6 We don't care about flow and...

Fri, 04 Jun 2010 21:09:50 +0000

Commit by mjagdis :: r5701 callweaver/corelib/connection.c: (link)

Make sure we initialize the whole of a sockaddr_in6

We don't care about flow and scope but they may be present.

Remove extra unlock in cw_context_create()

Fri, 04 Jun 2010 21:09:42 +0000

Commit by mjagdis :: r5700 callweaver/corelib/pbx.c: (link)

Remove extra unlock in cw_context_create()

$[ ... ]: Be generous in what we allow in strings This almost matches what is...

Fri, 04 Jun 2010 21:09:34 +0000

Commit by mjagdis :: r5699 callweaver/corelib/ (callweaver_expr2.l callweaver_expr2f.c): (link)

$[ ... ]: Be generous in what we allow in strings

This almost matches what is allowed during substitution and arg
splitting. Only almost though because we have to break when we hit an
operator. For consistent behaviour you HAVE to use quotes. (And they
have to work right :-) )

$[ ... ]: Make && and || short-circuit like C i.e. the second argument is only...

Fri, 04 Jun 2010 21:09:18 +0000

Commit by mjagdis :: r5698 callweaver/corelib/ (3 files): (link)

$[ ... ]: Make && and || short-circuit like C

i.e. the second argument is only evaluated if necessary. Essential if
the second argument is a function (bare, no ${...}) with side effects.