Re: Language support for flexible handling of system-detected errors.

["Marin David Condic" <dont.bother.mcondic.auntie.spam@[acm.org>]

Well, yeah, sure. If you have some interfacing capability, great. I guess I
was arguing against the idea that there could be one single standard out
there for all development tool interfaces or that any one given tool could
be designed to connect to a multitude of TBD other tools. Some simple
editors might provide no programming interface and no scripting language. If
you want to integrate them you'll have to do a lot of your own glue-code and
still not get everything you might desire. Some editors might provide a
single buffer while others provide many buffers. The editor might expect to
invoke other tools via its own scripts or it may provide a set of callable
interfaces - expecting to be called by something else - or both. Given the
multiplicity of ways that an editor might communicate (or not) with the rest
of the world and throwing on top of it the multiplicity of ways that
compilers, linkers, debuggers, GUI builders, code generators, analyzers,
configuration managers, etc., might possibly communicate with the rest of
the world, I don't think it is reasonable to expect some kind of standard to
emerge.

Where you do have specific tools and they provide some ability to
communicate, integrating them into a development environment is possible.
(Did this once in a previous incarnation. It was a really big job.) But
since there are so many possible ways the various tools could choose to
work, I don't think it is practical to hope for some kind of standard
interoperability. DEC did it their way - GNU does it theirs. (Standards are
such a wonderful thing that everyone should have one of their own! :-) Too
much of *how* it is done is going to be determined by the design of the
tools, so I think standards would be difficult to achieve. It may even be
undesirable because it might inhibit design choices when building the tools.

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com


"Stephen Leake" <stephen.a.leake.1@gsfc.nasa.gov> wrote in message
news:u8z871fga.fsf@gsfc.nasa.gov...
>
> Not to every editor on the planet, but there have been some efforts in
> this regard.
>
> DEC defined a standard format for compiler diagnostics (I forget what
> it was called), that LSE and other DEC tools could read. For non-DEC
> tools (like Emacs), there was a text equivalent.
>
> Gnu tools have a standard format for error messages, making them easy
> to parse by simple scripts in editors.
>
> Most programmer-oriented editors allow you to customize the command
> line used to invoke compilers and similar tools.
>
> I don't think we need a real "standard" here, just a common paradigm
> and a reasonably powerful scripting language in the editor.
>