Re: Auttomated tools

=2D----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello Dror,

On Monday 06 December 2004 1:41 pm, Dror Matalon wrote:
> Hi,
>
> Why does netWindows use its own documentation module rather than
> using doxygen?

'cause Doxygen requires significant cludges in order to be made to=20
work with JavaScript, whereas we can more directly model the way=20
things work in JS via DocBook. Mark Anderson went the doxygen route=20
with Burst, and I'm not sure I like what it does to the code. With=20
that in mind, we aren't doing our "own" doc system, we are=20
customizing a stock DocBook DTD and XSLT setup, and automating=20
generation of that documentation out of source code via Ant. If you'd=20
like help getting set up with the documentation portion of the build=20
system, let me know.

On a related note, I'm working on a hack to Rhino which will output=20
the documentation stubs that both the DocBook and Doxygen routes=20
still require. At that point, you should be able to pick your poison=20
via XSLT to a great extent.

> Also, has there been any attempts to use a javascript shell to
> find/catch errors in the code server side? Seems like automated
> scripts using Rhino or other javascript shells could catch syntax,
> and possibly other errors.

Short answer: yes, but it hasn't made it into the build system. Long=20
answer: Mark Anderson (a NW comitter) has built a toolkit=20
(http://burstproject.org) which provides most of these tools. The=20
problematic portions of NW testing have to do with browser-specific=20
behavior, of which there are many. MDA's answer was to emulate a=20
broad swath of this and build a unit-testing harness that works with=20
his browser-emulation setup. This isn't adopted by NW because of a=20
reliance on Perl.

> Ideally, there would even be a flag that would be equivalent to
> perl's use strict, or a 'C' lint type of check to find possible
> errors. I know about
> "http://www.crockford.com/javascript/jslint.html
> but somehow doubt that a script that doesn't know about the
> javascript syntax could do a through job.

Mark Anderson has also recently put togeather a JS "linker" which does=20
almost exactly this, and in running it against netWindows has exposed=20
several strictness warnings which I've tried to clean up. So from=20
that standpoint, it's another "yes, we have it, but it's not in CVS=20
or build.xml since it relies on Perl".

> This is not netWindow specific, but as I look at the code more and
> more, it seems like it's a necessity for a project this large.

I've been working on a new project (dojotoolkit.org) which implements=20
all of these things from the get-go, and I agree that it's helpful.=20
Back-porting much of this for netWindows isn't a huge priority right=20
now, though, but we'll run this stuff manually from time to time to=20
make sure we're sane.

Not a great answer for you, but hopefully it gives you a lay of the=20
land.

Regards

=2D --=20
Alex Russell
[email protected]   BD10 7AFC 87F6 63F9 1691 83FA 9884 3A15 AFC9 61B7
[email protected] F687 1964 1EF6 453E 9BD0 5148 A15D 1D43 AB92 9A46
=2D----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iD8DBQFBtPzooV0dQ6uSmkYRAh5fAKCnVJbyWxZ51e/KXUp6at5MJEFidQCfdSHM
nQKJkdu7iw+Y8dPtn11dOzo=3D
=3Dr61Z
=2D----END PGP SIGNATURE-----


_______________________________________________
The netWindows developers list: [email protected]
http://netwindows.org/mailman/listinfo/devel_netwindows.org