Re: libseed-list AlgoScript and custom module importer

["Alan Knowles" <alan akbkhome com>]

>From many years of working with the PHP core team, The one thing many people have noticed is be very wary of adding any kind of magic behavior it usually bites everyone in the backside eventually...

by the looks of things

algoscript = imports.algoscript;
myscript = algoscript.importFile('myalgo.ag');

is going to help everyone reading the code understand what it's doing

BTW = Once of the biggest issues with Seed at present, that I've not had time to look into is the lack of backtrace.. - this makes debugging a bit of a nightmare, print's everywhere ;)

This is probably going to be even worse with code builders unfortunately..

Regards
Alan


 --- On 11/Jun/2010, Jonatan Liljedahl wrote: 
> Robert Carr wrote:
> > On Thu, Jun 10, 2010 at 11:35 AM, Jonatan Liljedahl <lijon kymatica com> wrote:
> >> I'm working on a parser for "AlgoScript", which compiles to javascript and
> >> has some nice syntactic sugar and hides some of javascripts uglyness.
> >> Example:
> >>
> >>    [0..10].do {x -> print x}
> >>
> >> =>
> >>
> >>    (Array.range(0,10)).forEach(function(x){return(print(x));})
> >>
> >> I have an algoscript.parse() function that parses algoscript and returns a
> >> string of javascript.
> > 
> > This sounds cool! I sort of wanted to do this with coffeescript and Seed once.
> 
> I used the JS/CC with some minor modifications to support Seed, since 
> jison (used by coffeescript) had a lot of other js dependencies.
> 
> Now I have an imports.algoscript.Context class that works the same as 
> the sandbox, and uses the sandbox to run algoscripts in contained area. 
> It was very useful, since I can load my algoscript runtime library (some 
> prototype extensions) in this context instead of polluting the global 
> namespace.
> 
> I will put everything up on the net as soon as it's finished, would be 
> great to get some feedback!
> 
> >> One thing I'm thinking about is how I should solve module imports.. Surely I
> >> could have an import(name) function, but it would be a bit ugly to have this
> >> AND imports.gi.name for the gobject stuff.
> >>
> >> The best thing would be if imports.name could check the file extension and
> >> run it through my parser if it ends with '.as' or something like that. Any
> >> ideas how this could be achieved? Maybe the seed importer could have hooks
> >> that is called for files not ending with '.js'?
> >>
> >> algoscript_importer = function(filename,str) {
> >>  return eval(algoscript.parse(str));
> >> }
> >>
> >> imports.unknownFileCallback = algoscript_importer;
> >>
> >> Or it could just pass the filename and not the contents, or it could have a
> >> table of callbacks for each suffix:
> >>
> >> imports.customImporterCallback['as'] = algoscript_importer;
> >>
> >> What do you think? I imagine that this could be very useful for other stuff
> >> as well..
> >>
> >> I would be happy to try and add support for this if you want.
> >>
> >> /Jonatan
> > 
> > I'm a little hesitant with the file extension, but maybe some sort of
> > line to parse in the shebang? I can actually probably give a pass at
> > this in a few days, I'm getting ready to do a biggggg rework of Seed.
> 
> Actually, I don't really see the reason to ignore the file extension. 
> Strange things happen if you have foo.js, foo.txt, foo.html etc in the 
> module search path at the same time, this has bit me a couple of times. 
> Why would you name a javascript module to anything else than .js?
> 
> I would argue that parsing shebang would be more hackish (in a negative 
> sense ;) than to look at the filename extension.
> 
> What rework are you planning? Sounds interesting!
> 
> /Jonatan
> _______________________________________________
> libseed-list mailing list
> libseed-list gnome org
> http://mail.gnome.org/mailman/listinfo/libseed-list