From: Trans <transfire@...>
Date: 2007-09-05T19:40:07+09:00
Subject: Re: configuration file parser like Config::General for perl

On Sep 4, 1:21 pm, Tammo Tjarks <ta...@tammo-tjarks.de> wrote:
> Trans wrote:
>
> > On Sep 4, 11:35 am, Tammo Tjarks <ta...@tammo-tjarks.de> wrote:
> >> Hello,
> >> I have the problem, that in a project I am working in the configuration
> >> files are in a apache config file like format. This is a mixture of
> >> XML-Format and normal config format like
> >> <id>
> >>   name example
> >>   use  "for test"
> >> <id>
>
> >> The normal used language is perl, where the format, with some extensions,
> >> is supported by the Config::General Package.
> >> Is the something similar existing for Ruby? I have already searched in
> >> rubyforge, but I could not find something fitting. Somehow xmldigest or
> >> digestr seems partly to fit, but I am unsure if they support the whole
> >> range of the config-file. Or are there some settings existing for the
> >> xml-libraries like rexml to support this uncomliant kind of XML?
>
> > I know nothing of the sort. However, if should be pretty easy to use
> > REXML to loop through the XML elements and apply a parser to each
> > sections text.
>
> > T.
>
> Thanks for your answer. But I am not sure if that works without workaround.
> The Problem is, that the apache config file format allowas also things
> like:
> ---------------------
>
> version 1.0
>
> root "not existing"
>
> <display>
>  driver X11
>  name test
> </display>
>
> <modules>
>  perl 5.8.2
>  ruby 1.8.6
> </modules>
>
> --------------------
>
> I should mention that this is a meaningless example. The important point is,
> that you do not have proper xpaths and root-nodes. I have tried it once
> with rexml and it was somehow complaining (I have to look, but it was
> something with a line in rexml with <<)
>
> What I can try is to surround it by a master-node before read-in but that is
> a hack and I am not sure if it will work. Anyway I will try.

If they are only using Xml to do dead simple <x> ... </x> braketing,
then you could write a parser for that too. As long as there is no
nesting of identically named nodes, it is very simple parsing and
Regexp's can be used to do it.

Just a thought.
T.