> OK, I've implemeted the script to generate the extensions.xml file > automatically. > It's a quick&dirty script, but works perfectly :) > > The tags I've used are: > <!-- Purpose: xx --> > <!-- Membership: core, pecl, bundled, external --> > <!-- State: deprecated, experimental --> > > In the membership tag you can assign more than one attribute, separed by > commas. > When the script is run, it gives some debug info, like missing or bogus > tags. > > Let me know of any problem/suggestion.
Good start :) It would be much better to reintroduce the fixed TOC structure, since there were some intentions in the structure of the categorization (ie. basic, mostly used stuff was up in the TOC). Like text handling and variable and type handling should be above session, regardless of alphabetical order :) Two solution options: - parse the existing extensions.xml for structure - have a TOC strtucture array in the generator script Either way, bogus category tags can also be spotted this way. Goba
