[asterisk-users] ast_compile_ael2: Warning: file /etc/asterisk/extensions.ael, line 932-932: Empty Extension!

[Steve Murphy]

On Tue, 2008-06-03 at 17:18 +0200, Philipp Kempgen wrote:
> Steve Murphy schrieb:
> > On Mon, 2008-06-02 at 22:15 +0200, Philipp Kempgen wrote:
> >> On starting Asterisk (1.4) I get a whole bunch of
> >> WARNING[5858]: pbx_ael.c:4040 ast_compile_ael2: Warning: file /etc/asterisk/extensions.ael, line 932-932: Empty Extension!
> >> 
> >> I find it a bit disturbing that this message has a level of WARNING
> >> (instead of NOTICE maybe) because the extensions in question are
> >> empty on purpose. The only reason they exist are the hints.
> >> 
> >> hint(SIP/3000) 3000 => {}
> >> hint(SIP/4000) 4000 => {}
> >> ...
> >> 
> >> All the "real" dialplan stuff is in
> >> _Z. => {
> >> 	// do something ...
> >> }
> >> 
> >> Maybe the error level should be lowered so the user is not mislead
> >> into thinking something was broken?
> >> Or is there a way to do it which does not trigger the warnings?
> 
> > I understand your frustration.
> 
> Frustration is not the right word, I'm not frustrated.
> 
> Telling from the zero feedback from other users to my post what I
> do seems to be rather uncommon. So I wanted to make sure that
> a) there is no better way to do it and b) that I'm not doing
> something which is considered deprecated and subject to stop
> working sooner or later.
> 
> It's just that people see my dialplan causing lots of warnings so
> they assume it's broken.

Philipp--

I wouldn't worry about why users don't respond to a message. It's
a high-volume mailing list, and people are sometimes too busy to
answer. Such stuff only gets a response if a high emotional response
is evoked, and your request might not have rated right at this part
of the day. Or... nobody but you is using AEL! (which I somewhat doubt).

> 
> > So, my code prints a warning, and fills in the priority with a call
> > to NoOp().
> > 
> > Now, in the end, it doesn't matter at all what the internal
> > representation is,
> > or what the package is doing deep inside.
> > 
> > So, unless users in general object, I will go ahead and just remove
> > this warning. What a compiler has to do underneath to make your 
> > code work/palatable is not necessarily of interest to you.
> 
> Probably I could use
> hint(SIP/3000) 3000 => {NoOp(make the compiler happy);}
> as long as _Z. matches before the literal extension does.
> 
> Changing the warning to a notice would do for me. Removing it
> is fine as well - as long as _you_ don't have any objections. :-)
> 

I had so little objection, I committed the discussed fix
to 1.4 with rev. 119929
to trunk with rev. 119930
and to 1.6.0 with rev. 119931

and then I made a pass to update the regressions for ael and committed
those to 1.4 with rev. 119966
and to trunk with rev. 119998
and to 1.6.0 with rev. 120000

(hey! a nice round rev number!)


> 
> Grüße,
und du, auch!

> Philipp Kempgen
-- 
Steve Murphy
Software Developer
Digium
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3227 bytes
Desc: not available
Url : http://lists.digium.com/pipermail/asterisk-users/attachments/20080603/980b7ce8/attachment.bin