Re: Creating IDE for internal script language (George C. Lindauer)
10 Jul 1996 12:01:15 -0400

          From comp.compilers

Related articles
Creating IDE for internal script language (1996-07-09)
Re: Creating IDE for internal script language (1996-07-10)
| List of all articles for this month |

From: (George C. Lindauer)
Newsgroups: comp.compilers
Date: 10 Jul 1996 12:01:15 -0400
Organization: University of Louisville, Louisville KY USA
References: 96-07-060
Keywords: tools writes:

>I am working on an IDE for an internal script language. I have
>identified several problem domains I am not familiar with
>and the biggest one for me is how to:
>a.) have the IDE be able to create a script from the user's
>editing of graphical objects on screen.

This is heavily domain-specific. Most parsers generate that same type
of internal data structures that your IDE is probably going to have to
create to keep track of the graphical relationships. So to continue
to use the original parser you are going to have to go backwards... create
the script from the type of internal representation the parser uses. This
is the same thing compilers do in the code generation phase, only you will
be working at a higher level than assembly.

>b.) let the user hand edit the script generated from a.) if
>hand customization is needed

Once you've got the script this is easy, you just have to have an editor.
You can grab one from the public domain and modify it, you can write your
own, or you can simply shell out to your favorite editor.

>c.) be able to compile either a.) or b.) with our parser

Once you've got the script you just shell out and run the parser... or since
you have the sources you can integrate it into your IDE.

>d.) have the IDE recognize the output of a.) or b.) without
>reimplementing the parser over again and display the elements
>in a.)

You pretty much have to have a parser for this. You can get your script
generator to put comments and hints in and use that... except you have
said you might be hand-modifying the code.

>The complication is that someone else wrote the monolithic
>parser we are using and there are no intermediate files, it
>goes directly to binary from there in the yacc script.

What I would do is to examine the parser you have and determine what kind of
internal representation it uses. Then use something as similar as possible
for the IDE. With a little work you'll be able to use that same parser
both to get the IDE representation you need and the binary code you need.
With this approach, the only major code to be written is the script
generator (and all the gui stuff you are probably doing).


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.