RE: Fun with Garnet & ASDF - Garnet ASDF cursor system, defining cursors at system load time

Hi Sean,
Thanks for this. Does it make sense for files like this to be added to the SourceForge
Garnet area or something so they aren't lost?
http://garnetlisp.sourceforge.net/
http://sourceforge.net/projects/garnetlisp

Brad Myers

> -----Original Message-----
> From: Sean Champ [mailto:[email protected]] 
> Sent: Monday, August 29, 2005 5:24 PM
> To: Garnet Users List
> Subject: Fun with Garnet & ASDF - Garnet ASDF cursor system, 
> defining cursors at system load time
> 
> Hello,
> 
> In order to make it somehow easier to load and develop 
> Garnet, and to incorporate Garnet with other Common Lisp 
> applications that do and would have ASDF system definitions, 
> I've been taking some time to define ASDF systems for Garnet. 
> I hope to share the work, made for this, once it would be all 
> completed and tested -- I've got as far as defining a system 
> for Garnet-Gadgets, then had to stop to address how I will 
> have GARNET-BITMAP-PATHNAME defined, given how I've been 
> doing the Garnet ASDF "stuff", thus far. At that time, I 
> stopped to notice the Garnet cursor files.
> 
> The ASDF work, mentioned, is not my primary point, here, 
> though it has been coincident to this email. When it would be 
> released, I hope to write enough notes to explain it.
> 
> 
> 
> As a result of the ASDF+Garnet fun, I decided to define 
> something, upon a lark, about Garnet cursor files. I thought 
> it would be worth sharing, then. It's attached -- an ASDF 
> system definition, containing some  extensions on ASDF; it 
> would best be located within the Garnet lib/bitmaps/ 
> directory, named garnet-cursors.asd, if one would load it. It 
> would be most easily used, if one would already have Opal 
> loaded. Then, one would simply load the file, and execute
> 
>   (asdf:operate  'asdf:load-op :garnet-cursors)
> 
> Of course, it would redefine some things, but just as how 
> they've been defined, before.
> 
> 
> Essentially, the code in the attached ADSF file does 
> represent a means for automatically defining Garnet cursors 
> when an ASDF system is loaded -- such that the system would 
> contain one or more CURSOR-COMPONENT components. I'm sure it 
> is not perfectly  done, but it may at least serve as an 
> example, and a first revision (avaialble in later forms, once 
> it would be availabile)
> 
> 
> As for  how the attached ASDF extensions would work, I'll 
> explain it in regards to a hypothetical ASDF component with 
> the name <FOO>. This explanation might be more clear, after a 
> brief look at the attached *.asd file.
> 
> 
> I'd thought that this set of extensions on ASDF would be 
> helpful, for that it would allow a Garnet developer to simply 
> create the appropriate <FOO>.mask and <FOO>.cursor files, 
> then identify a <FOO> CURSOR-COMPONENT in an ASDF system -- 
> with no more code being required, for this.  When the ASDF 
> system would be loaded, then, the apprpriate <FOO>-CURSOR and 
> <FOO>-CURSOR-MASK schemas would be defined for the <FOO> 
> CURSOR-COMPONENT; both of those schemas would then be 
> identified within a <FOO>-PAIR variable -- much like how it 
> is being done in Garnet, though this would do it 
> automatically, upon the loading of the ASDF system.
> 
> Additionally, this allows for a :PACKAGE argument to be 
> specified in the definition of the comopnent; it would be 
> spedified in the ASDF definiton for the <FOO>  
> CURSOR-COMPONENT, or in the definiton of the module (possibly 
> a system) directly containing the <FOO> component. Granted, 
> the attached *.asd file is not so well desinged for this -- 
> some class names may need to be revised -- but it works.
> 
> The :PACKAGE identified will have to exist when the component 
> would be loaded; it would be used as the package in which the 
> symbols for the CURSOR-COMPONENT would be interned (namely, 
> the cursor schemas and the cursor's <FOO>-PAIR variable)
> 
> 
> Granted, this serves to require that one wold be using ASDF, 
> to load the appropriate system. I'm sure that that should not 
> be a matter of problem; ASDF is hardly an uncommon application.
> 
> 
> Furthermore, now having completed the thing, I've come to 
> realize that this particular system definition, attached, it 
> is not so glaringly necessary; there is already the code, 
> within Garnet, for acccomplishing what the attached system 
> definition does accomplish.
> 
> Still, if anyone would be defining new cursors for Garnet, 
> and using ASDF, then perhaps the attached code would be of interest.
> 
> 
> I hope it serves to demonstrate some of what can be done with 
> ASDF systems, furthermore. 
> 
> Something like this could be done, if necesary, for 
> automatically loading Garnet bitmaps, when an appropriate 
> ASDF system would be loaded. Of course, it could be done for 
> any image file, given some appropriate support in the Lisp image.
> 
> (Incorporating Garnet with more of what exists for the 
> representation of image files -- I do not know the 
> package-names offhand, but I know they exist, and they should 
> be at http://common-lisp.net/ -- it  is one more item on my 
> own to-do list; perhaps there will be come code after it, 
> sometime when.)
> 
> Incidentally, the ASDF+Garnet work is being done as part of 
> an intention, for to package Garnet for Debian GNU/Linux 
> systems. That should be reported about, when it's done and tested
> 
> I do not intend to work-in this garnet-cursor code with it, 
> except perhaps as an example. Regardless, it's done, and 
> attached, in at least a first draft form. Maybe it will have 
> been of any interest.
> 
> 
> 
> 
> --
> Sean Champ
> [email protected]
>