[gnome-love] Re: approval from sun and a last XML request [was Re: [gnome-love] wanted: php or cgi programmer with some spare time to help with gnome test spec]

[Luis Villa <louie ximian com>]

BTW, I think this has reached the point where it should really move off
of gnome-love and on to gnome-web-devel, both to reduce irritation on
the part of the gnome-love people and perhaps to pick the minds of the
gurus on gnome-web-devel :)

On Mon, 2001-12-03 at 11:01, fflewddur wrote:
> On Mon, 2001-12-03 at 10:00, Luis Villa wrote:
> > BTW, just wanted to let everyone know that I've talked to the current
> > maintainer of this project at Sun and he's very open to using this code.
> awesome.
>
> > Todd: question: if I drop a new file into one of the subdirectories,
> > will this page
> > http://www.shopindiego.net/gnome_test_specs/components/battery_charge_monitor_applet/
> > get updated?
> no.  i was confused.  i thought you just wanted the ability to drop in a
> new module, not add a new test to an existing module. 

Well, ideally, both. The idea is to make it as easy as possible for
maintainers and/or QA volunteers to add their own stuff. The most
relevant example are regression tests: it's good QA practice to add a
new test for each bug that is 'fixed' to make sure it doesn't recur;
adding those in any way other than one at a time is a PITA.

> thanks to the
> perl interface, however, changing this shouldn't be more than a
> one-night job. 

I was hoping not :)
 
> > > i couldn't figure out where the
> > > instruction html files are generated (or if they are at all...), so
> > > there's still a lot of cruft there that could go. 
> >
> > AFAIK, they aren't- I think it's a template page that they fill in and
> > save. Before calling this a day and starting to push this into CVS, is
> > anyone still open to the idea of converting the test files to XML or
> > simple text, so that we can easily change their presentation?
> > Luis
> well, i tracked down three perl scripts that are *supposed* to generate
> the instructions, and while they do run during the make_all_indices
> script, they didn't actually change anything :P

Which ones are those? I'll take a peek.

> i'm not very familiar with xml yet, but i've been meaning to learn. 
> what exactly would be needed in order to move this whole thing over to
> xml files?

1) An XML format; wouldn't need to be more complicated (at first) than
<test>
<component></component>     }- these two might be optional- i.e., if
null/not present, then just build them from the dir structure?
<subcomponent></component>  }
<testcasename></testcasename>
<testaim></testaim>
<procedure></procedure>
</test>
but this is just a suggestion; I'm sure some XML gurus on the list will
have a better idea.

2) a perl script to convert the current HTML into that format

3) some php magic to read the XML and convert it into something
prettified for the web. Since this new HTML would be dynamically
generated, I figure it would be easy to edit/make prettier/prune useless
info/etc. Also, since the data is in a nice, parseable format, I'd hope
it would be trivial to add the <testaim> field to the test page, and
thus dispense with the table of contents page altogether (maybe.) 

Luis


-- 
Luis Villa
Ximian Bugmaster
"Quality is an amazing bridge because it is universal in its language."
Thomas Corcoran