Re: [Geiser-users] eval commands always split debug window?

[Scott Messick]

> >  For commands with no output, they display a result in the minibuffer. 
> > For commands with output, whether or not there is an error, they split a
> > debug window and display their results and output there.  Is this
> > behavior intended?
> 
> yes.  the debug window is the recipient of output.  due to
> implementation reasons (the communications between the repl process and
> emacs use the process's sttandard input/output), the output of an
> evaluation cannot be shown in the repl buffer.
> 
> > Regardless, I'd like to change it so that these commands just send their
> > contents to the REPL.
> 
> sorry, but i'm afraid this is not currently possible (see above).

Okay, good to know.

> i'm not sure i understand correctly, but are your commands writing their
> results to stdout and that's why you need syntax highlighting? when you
> simply evalutate a scheme value, without writing to stdout (which is the
> regular operation for me), the evaluated value appears in the
> mini-buffer with syntax highlighting, and also in the debug window with
> syntax highlighting (but the window is not shown most of the time).
> 
> in my case, it's very rare that my evaluations write to stdout (i.e., i
> almost never evaluate things like (display foo) or (write bar)), but
> certainly YMMV.

Yes, I do have some things writing to stdout, mainly for
debug/development purposes.  However, the bigger issue when I evaluate
something that causes errors.  The lack of syntax highlighting makes the
trace much harder to read., and because it's long I'd prefer not to have
it in a half window.

I have not really established my workflow here, so I think all of this
is fine and I can figure out what I really want.  On reflection, a lot
of the trouble comes because in what I'm working on now, I occasionally
have top-level definitions of data (like a list or hash-table, not a
function).  So code mistakes would be likely to produce errors
immediately when these are evaluated, whereas functions would be defined
okay (with C-c C-c) and then signal errors when I test their use (in the
actual REPL buffer).  Probably I can fix this by adjusting my coding
and/or debug style a little bit.

Thanks for your help!

Scott

On Fri, Feb 24, 2017, at 23:21, Jose A. Ortega Ruiz wrote:
> On Fri, Feb 24 2017, Scott Messick wrote:
> 
> > I'm trying to start using Geiser with Chicken (new to both, but I have
> > used SLIME in the past and very briefly used Geiser with Racket).
> >
> > The eval commands such as C-c C-c or C-c C-b seem to never use the REPL.
> 
> they are not designed to use its buffer.  the repl's process is of
> course used to perform the evaluation, and obtain their results.
> 
> >  For commands with no output, they display a result in the minibuffer. 
> > For commands with output, whether or not there is an error, they split a
> > debug window and display their results and output there.  Is this
> > behavior intended?
> 
> yes.  the debug window is the recipient of output.  due to
> implementation reasons (the communications between the repl process and
> emacs use the process's sttandard input/output), the output of an
> evaluation cannot be shown in the repl buffer.
> 
> > Regardless, I'd like to change it so that these commands just send their
> > contents to the REPL.
> 
> sorry, but i'm afraid this is not currently possible (see above).
> 
> > Is there a simple way to do this?  The debug windows, apart from being
> > annoying, also seem useless: lack of syntax highlighting makes them
> > hard to read.  I'm not sure if that's just an issue with my setup,
> > though.
> 
> i'm not sure i understand correctly, but are your commands writing their
> results to stdout and that's why you need syntax highlighting? when you
> simply evalutate a scheme value, without writing to stdout (which is the
> regular operation for me), the evaluated value appears in the
> mini-buffer with syntax highlighting, and also in the debug window with
> syntax highlighting (but the window is not shown most of the time).
> 
> in my case, it's very rare that my evaluations write to stdout (i.e., i
> almost never evaluate things like (display foo) or (write bar)), but
> certainly YMMV.
> 
> and certainly i might be misunderstanding :)
> 
> hope this helps a bit,
> 
> cheers,
> jao
> -- 
> If a listener nods his head when you're explaining your program, wake
> him up.
>   - Alan Perlis, Epigrams on Programming
>