| Newsgroups |
gmane.mail.exim.vexim.devel |
| Message-ID |
<[email protected]> |
Hi,
Le 23 juil. 08 à 00:31, Tim Aslat a écrit :
> A re-write may well be on the cards, but I would suggest re-
> building the
> core in a slightly more modular format, without changing the language.
>
> My main reason for using vexim was the fact that it was written in
> PHP,
> making my learning curve so much easier as I already have enough on my
> plate maintaining a whole stack of Perl, PHP and some custom scripting
> systems, without having to learn another language.
Why not split vexim ? the specific and complex part is exim4
configuration and association with antispam
filter and sql db tables. Many frameworks could be used to handle
lists of users and options.
That's why I would propose to split vexim into vexim-core, which
would have a conf file to tell it the name of tables, fields, and all
options (which spam filter, which version of what) and would have a
good documented api. All this can ba achieved by a bash or tclsh
script and a man page.
Maybe this command line tool could even generate exim4 config files.
Then this rock solid core could be used by many users to build their
own language specific system with gui, in their preferred language
(php5, python, ruby and so on).
Jean-Christophe Michel