[asterisk-dev] bugs a plenty - discuss....

[Luigi Rizzo]

On Thu, Mar 09, 2006 at 01:34:17PM -0600, Kevin P. Fleming wrote:
> Paul Cadach wrote:
> 
> > Don't forget about my ticket which splits huge say.c into many small
> > language-dependent pieces and provides common interface to support different
> > languages. Also, next step was planned is to move some language-specific stuff
> > into it like replace channel's "char *lang" variable with pointer to language
> > structure to not perform many string comparsions each time.
> 
> I think it sounds like Paul and Luigi have good, but differing, ideas on
> the best way to handle this. I also see no need to make any 'interim'
> steps; we are still in 1.4 'architectural change' mode until the end of
> this month, so let's get this figured out.
> 
> I think it does make quite a bit of sense to make the approach
> table-driven as Luigi has proposed, but we need to ensure that we don't

note that the credit for the idea goes to crich, i just followed up.

> lose the ability to handle the complexities of languages with unusual
> ways of speaking numbers, dates and times. Using C code allows for that,
> but at a cost.

we can get that in the table- by providing some escape strings
in the 'rules' that result in calling the specific C code
(besides, we can do very convoluted string manipulation
using the variables and pattern matching so i am confident
that the approach can cover everything very easily).

> (Also, I'm happy to see these discussions moved back to this list...
> Mantis is _not_ a good place for architectural or design discussions at
> all).

so how about an interim solution where we carry on both the old say.c
and the new app_say.c, with a config-file switch to choose which
one, and after a while when we are confident that the new methods
is on par with the old one we get rid of say.c and things
keep working as before (or hopefully better)

kevin ? mark ?

	cheers
	luigi