My personal site (the one linked to over on the right) has been nagging at me for quite some time. I wanted to make it easier to find stuff, and also to have autogenerated "last mod" times on things that should have them. Tonight I finally sat down and banged out a really ugly solution that does a fairly good job for a first whack.
One thing I *didn't* want was a CMS; I wanted an intelligent index page generator. And I didn't want dynamic (that is: via apache) generation; I wanted something I could call from the shell script that rsyncs my websites, and have it all happen at once.
So I sat down and whipped up a mini-markup language and an exceedingly stupid parser to handle it, /et viola/, new homepage which doesn't involve typing any HTML.
It can do headers, file import, lists (all 3 HTML types), tables, links, and escape its special characters. It generates validated, properly-nested XHTML1.1, and the parser weighs in at 5K of very ugly, uncommented code. Oh, and its markups can be nested due to abusing $_ and reading from filehandles that are controlling loops in new and exiting ways :D
It is, due to the vagarities of my project naming scheme, called "egg", and it's the last thing the world needs, but it was fun.
Textile is meant to do similar things. There's also a perl module out there to do the conversion.
Re:Non-HTML markup...
mdxi on 2003-11-21T15:30:32
Now that you point me at it, I remember seeing that some months ago. I wanted a little more automation on certain things, though, as this is a one-trick pony instead of a general-purpose markup thinger.
I do have one of those coming though.
