Re: [Geiser-users] Help associating source and REPL - "No prompt found!"

[Jon Miller]

I really want to get this working, so I went ahead and launched edebug and
here is my mediocre notes that I took during the session:
; Using edebug-defun on geiser-debug--send-region
;  gets stuck on (ret (geiser-eval--send/wait code))
;    looks like it is (geiser-eval--connection) that yields the missing REPL 
connection message
;      that comes down to geiser-repl--connection which relies upon 
geiser-repl--connection*
;        Looks like the assigned buffer name doesn't work. 
(geiser-repl--set-up-repl geiser-impl--implementation) yield nil
;          Would have set impl to 'racket' had (geiser-repl--repl/impl impl) 
returned non-nil
;            Within geiser-repl--repl/impl, geiser-repl--repls is nil and repls 
isn't passed, so the dolist loops through nothing.

I'm not sure if the variable geiser-repl--repls being set to nil is the key
in this analysis or not, but it would appear so to me. I looked at it's
initial definition but haven't yet see where it could be updated. 

Perhaps I'll step through run-geiser after some lunch. 

-- 
Jon Miller

address@hidden writes:

> address@hidden writes:
>
>> Hi Jon,
>>
>> On Wed, Nov 13 2013, Jon Miller wrote:
>>
>>> Hello Geiser folks,
>>>
>>> I've just installed Geiser last evening and plan on using it while
>>> working through the SICP book with folks from work. Loving it so
>>> far. The other thing I'm trying to do is to incorporate it with
>>> literate notes within a org-mode file[1]. I'm having problems getting
>>> a geiser REPL running and communicating with that from my org file.
>>>
>>> I believe I've narrowed down the problem to a problem associating my
>>> source file with the geiser REPL. With a plain scheme file
>>> example.rkt, I have the example code: (define (cube x) (* x x x))
>>>
>>> When I run M-x run-geiser, it will start a REPL and split my window
>>> but I'll get a message in the minibuffer "No prompt found!" In my
>>> *Messages* buffer, I'm seeing the following:
>>> geiser-repl--wait-for-prompt: No prompt found!
>>
>> That typically happens when the scheme interpreter version is not the
>> right one.  What versions of Racket and Geiser are you using? You can
>> also try M-x run-racket to make sure that you're starting Racket.
> I get the same behavior with M-x run-racket. (I've set
> geiser-active-implementations to '(racket) to avoid guile).
>
> Version of Racket is: v5.3.6. I installed it directly from the Racket website
> to my Fedora 18 machine.
> M-x geiser-version says 0.5. I installed it via an el-get recipe which looks
> like it does a git clone from git://git.sv.gnu.org/geiser.git.
>
>>> I've tried playing with various values for geiser-repl-startup-time
>>> from the smallest 1000 to the largest value of 20000. Same result each
>>> time.
>>>
>>> I later try M-x scheme-send-definition as well as M-x
>>> scheme-send-region and the simple cube definition never makes it's way
>>> to my REPL.
>>
>> No, those functions are not part of Geiser, and you shouldn't be using
>> them (they're part of the standar emacs support for scheme that Geiser
>> supersedes, and won't work for a REPL started by Geiser).  Once you get
>> Geiser up and running, there are replacements for them that you can use.
> Okay, I'll be careful about which I use. Atleast, C-x C-e is bound to
> geiser-eval-last-sexp but again once I run that after a definition, I get
> "geiser-repl--connection: No Geiser REPL for this buffer (try M-x
> run-geiser)" in *Messages*. Each time I run M-x run-geiser, I keep getting a
> separate REPL instantiated.
>
>>> So, I'm hopefully providing enough information and you can see why I
>>> might be concluding a communication and/or problem properly launching
>>> the Geiser REPL is likely my root problem. Any suggestions?
>>
>> That seems to be the problem, yes.  But i don't know how ob-scheme.el
>> works, so i'm not totally sure.  At any rate, if you cannot start a
>> Geiser REPL, it's almost for sure *not* ob-scheme's fault.
> I think we can ignore ob-scheme for now since I suspect if I can solve the
> basic scheme source <--> REPL problem, I suspect I can figure out the org
> file setup on my own.
>
> Is there a debug level I can set within Geiser? Or would you recommend I step
> through with Edebug? (still kind of a novice there)