Using a pure bash XML "parser" for jhlfs. A very humble suggestion awaiting for comments.

George Makrydakis gmakmail at gmail.com
Fri Jan 27 13:43:05 PST 2006


Jeremy Huntwork wrote:
> Very nice, George. I've done something similar in C (it only depends on
> standard libs). I never got around to finishing it, though. :/ The
> advantage of Bash, would be that hopefully, more would be able to help
> out. To me, I've found that the hard part is iterating through all the
> xi:include tags in the LFS book. Also, keeping track of and swapping out
> the entities and their values can be difficult.
> 
> Anyway, thanks for this. If you get it to a point where it does what the
> current XSL script does, we just might be able to drop it in. :) Fewer
> dependencies are always nice.
> 
> --
> JH

Thank you for your response.

I am working on such a solution, again using pure bash (no other 
binaries involved, whatever those may be). It is possible to have such 
an "iterator", that is also relatively fast for a script. I do have a 
working (but buggy) prototype for this.

The problem can be solved if you build other scripts that are capable 
of handling complex data structures (ie "bi-dimensional" arrays --> 
tables) in bash without patching it.

Provided I have the time to work on it the next few days you will be 
definitely seeing something of interest regarding XML processing 
commands, comment support and _fast_ entity dereferencing.

So, soon to come, just for the fun of it!

gmak




More information about the alfs-discuss mailing list