[Scilab-Dev] Re: and idea for the development of Scipad

[Sylvestre Ledru]

Le samedi 18 octobre 2008 à 11:04 +0200, Torbjørn Pettersen a écrit :
> François Vogel skrev:
> >> By the way, it would be great to have
> >> unitary tests here.
> >
> > You mean, for function help_from_sci, right? Not for the feature in 
> > Scipad I guess.
> >
> > Also an updated help file would be needed. The one I currently have is 
> > consistent with the very first version of help_from_sci. It does not 
> > deal with the new third argument.
> >
> > Both points left to the author of help_from_sci unless he really does 
> > not want to go ahead.
> With "both points" you mean unitary tests and a new helpfile, right?
> 
> An updated helpfile is enclosed. Writing documentation seems to kill my 
> writing skills, so if I've missed or miss explained something just let 
> me know.
> 
> But what is meant with unitary tests? A set of test cases to check the 
> validity of the code? Are there any guidelines for writing such tests?
As François said, the documentation is not really up-to-date here... It
is not because we have been lazy but especially since writing tests is
pretty easy. The easiest way to do understand it is read other tests
like in :
modules/helptools/tests/unit_tests/xmltopdf.tst
Basically, for you, this would be:
- First test:
* create a fake scilab function with comments
* save it in a file
* launch your function on it
* check that the output is, at least, not null (and contains good stuff)

- Second test:
* create a fake function wit bad comments and redo the process (to check
that your function won't crash if the comments are badly formated).

> By the way - what is SEP an acronym for? I'm usually good at deciphering 
> three-letter-abbreviations, but not this one.
Scilab Enhancement Proposal

> You may want to check that the links in the help_from_sci help file 
> works when including it into scilab.
> help_from_sci generate links to the help pages of the functions listed 
> in the "See also" section of the help file.
> The xml code looks like this:
> 
>     <refsection>
>        <title>See also</title>
>        <simplelist type="inline">
>        <member><link linkend="help">help</link></member>
>        <member><link linkend="help_skeleton">help_skeleton</link></member>
>        <member><link linkend="xmltohtml">xmltohtml</link></member>
>        </simplelist>
>     </refsection>
> 
> which as far as I can tell should be consistent with the xml format used 
> in Scilab 5.
yep, that is the way to do it.

Thanks again for your work!
Sylvestre