I played around with mscorlib.xml today. Enclosed are some files to
generate a really raw set of index pages.
On Sat, 2002-10-12 at 15:59, Miguel de Icaza wrote:
> Well, the various "scripts" should be used to generate different
> presentations (class info, complete class dumping, specific member,
> member summary). So just assume that you can jump to all of those entry
> points.
OK, right now they're generating static html files, locally browsable.
All the information needed to generate the files is contained in the
link, i.e.
> I would like to only ship the XML files, and have our documentation
> browser generate the various presentations on the flight using the XSLT
> transformations.
It takes both Sablotron and xalan C++ the better part of a second just
to read in mscorlib.xml (Athlon 1800, 512 Mb RAM). Breakdown: parsing
the input files takes 900 ms, the transform takes 150 ms. I can cut
this down by doing a pre-transform and caching the results.
Where did mscorlib.xml come from? Can we add class summaries to it?
Right now, if I want to create a page that lists classes and their
summaries
> > 3) Where can I find the other XML files? Is there a DTD or spec
> > or something? I've only tested against Types.xml.
>
> I have now checked in all the XML files into the Mono CVS repository in
> the module `monodoc' (I had forgotten to push it to the anoncvs servers,
> but am forcing a push right now).
#!/usr/bin/perl -w
my $xml = 'mscorlib.xml';
open(FH, "xalan -Q -XSL list-namespaces.xsl -IN $xml |")
or die "Could not list namespaces: $!\n";
my @arr = <FH>;
close FH;
print "Generating (root)\n";
system("xalan -Q -XSL show-namespace.xsl -IN $xml -OUT show-namespace.xsl--[].html") == 0 or die "Could not run xalan: returned $?\n";
for(@arr)
{
chomp;
print "Generating $_\n";
system("xalan -Q -XSL show-namespace.xsl -IN $xml -PARAM ns \"'$_'\" -OUT show-namespace.xsl--[$_].html") == 0 or die "Could not run xalan: returned $?\n";
}
