From nikodemus at random-state.net  Sat Jun  7 12:52:59 2008
From: nikodemus at random-state.net (Nikodemus Siivola)
Date: Sat, 7 Jun 2008 15:52:59 +0300
Subject: [climacs-devel] Some Climacs issues
Message-ID: <633d72b0806070552w1a389abfl3f6d5a88d78f9f71@mail.gmail.com>

1. C-x C-f foo.lisp, then Menu: Help->Describe Bindings

Window splits into two views, and the Lisp view remains the active
one. C-x C-b switches to the Help view, but the only visual clue that
it is indeed the active one is that the cursor in the Lisp view turns
blue. If there are multiple cursorless views visible, it is impossible
to tell which one is currently active -- at least as far as I can
tell.

2. Lack of cursor-motion keybindings in the Help view is another
thing, but I'm sure you knew that one. :)

3. Finally, mousing over the binding names in the Help view leaves the
focus boxe visible after the pointer moves over another binding name.
Scrolling clears the focus boxes.

If one or more of these are likely to be simplish to fix, I would not
be averse to hints as to where to look.

Cheers,

 -- Nikodemus


From athas at sigkill.dk  Sat Jun  7 13:07:25 2008
From: athas at sigkill.dk (Troels Henriksen)
Date: Sat, 07 Jun 2008 15:07:25 +0200
Subject: [climacs-devel] Some Climacs issues
In-Reply-To: <633d72b0806070552w1a389abfl3f6d5a88d78f9f71@mail.gmail.com>
	(Nikodemus Siivola's message of "Sat, 7 Jun 2008 15:52:59 +0300")
References: <633d72b0806070552w1a389abfl3f6d5a88d78f9f71@mail.gmail.com>
Message-ID: <874p85o0ky.fsf@lambda.athas.dyndns.dk>

"Nikodemus Siivola" <nikodemus at random-state.net> writes:

> 1. C-x C-f foo.lisp, then Menu: Help->Describe Bindings
>
> Window splits into two views, and the Lisp view remains the active
> one. C-x C-b switches to the Help view, but the only visual clue that
> it is indeed the active one is that the cursor in the Lisp view turns
> blue. If there are multiple cursorless views visible, it is impossible
> to tell which one is currently active -- at least as far as I can
> tell.

I've been thinking about this now and again, and there are many ways
to fix it I think. One idea I've had is to encapsulate each window in
a border pane and disable/enable the borders when the active window
changes.

> 2. Lack of cursor-motion keybindings in the Help view is another
> thing, but I'm sure you knew that one. :)

This is hard to fix, because there isn't a cursor at all in the Help
view.

> 3. Finally, mousing over the binding names in the Help view leaves the
> focus boxe visible after the pointer moves over another binding name.
> Scrolling clears the focus boxes.

Yeah... this bug exists in the first 150 lines of climacs/typeout.lisp
or so and is related to the great lengths Climacs go to in order to
make typeout views transparent for the programmer. I'm not sure how
easy it is to fix, but it's not likely to require much code.

-- 
\  Troels
/\ Henriksen


From nikodemus at random-state.net  Sat Jun  7 13:23:11 2008
From: nikodemus at random-state.net (Nikodemus Siivola)
Date: Sat, 7 Jun 2008 16:23:11 +0300
Subject: [climacs-devel] Some Climacs issues
In-Reply-To: <874p85o0ky.fsf@lambda.athas.dyndns.dk>
References: <633d72b0806070552w1a389abfl3f6d5a88d78f9f71@mail.gmail.com>
	<874p85o0ky.fsf@lambda.athas.dyndns.dk>
Message-ID: <633d72b0806070623u199d75b6ked08b0df3273e259@mail.gmail.com>

On Sat, Jun 7, 2008 at 4:07 PM, Troels Henriksen <athas at sigkill.dk> wrote:

> I've been thinking about this now and again, and there are many ways
> to fix it I think. One idea I've had is to encapsulate each window in
> a border pane and disable/enable the borders when the active window
> changes.

Borders would be great.

>> 3. Finally, mousing over the binding names in the Help view leaves the
>> focus boxe visible after the pointer moves over another binding name.
>> Scrolling clears the focus boxes.
>
> Yeah... this bug exists in the first 150 lines of climacs/typeout.lisp
> or so and is related to the great lengths Climacs go to in order to
> make typeout views transparent for the programmer. I'm not sure how
> easy it is to fix, but it's not likely to require much code.

Semi-relatedly: what is the best way to discover what kind of an
object a given view is? Is there a way to make right-click offer me a
"Descibe Pane" or something like that for any and all panes?

Cheers,

 -- Nikodemus


From athas at sigkill.dk  Sat Jun  7 13:29:47 2008
From: athas at sigkill.dk (Troels Henriksen)
Date: Sat, 07 Jun 2008 15:29:47 +0200
Subject: [climacs-devel] Some Climacs issues
In-Reply-To: <633d72b0806070623u199d75b6ked08b0df3273e259@mail.gmail.com>
	(Nikodemus Siivola's message of "Sat, 7 Jun 2008 16:23:11 +0300")
References: <633d72b0806070552w1a389abfl3f6d5a88d78f9f71@mail.gmail.com>
	<874p85o0ky.fsf@lambda.athas.dyndns.dk>
	<633d72b0806070623u199d75b6ked08b0df3273e259@mail.gmail.com>
Message-ID: <87zlpxmkz8.fsf@lambda.athas.dyndns.dk>

"Nikodemus Siivola" <nikodemus at random-state.net> writes:

> Semi-relatedly: what is the best way to discover what kind of an
> object a given view is? Is there a way to make right-click offer me a
> "Descibe Pane" or something like that for any and all panes?

For my own hacking, I use this set of commands:

(define-command (com-inspect-buffer :name "Inspect Buffer" :command-table base-table) ()
  (clouseau:inspector (current-buffer)))

(define-command (com-inspect-frame :name "Inspect Frame" :command-table base-table) ()
  (clouseau:inspector *esa-instance*))

(define-command (com-inspect-window :name "Inspect Current Window" :command-table base-table) ()
  (clouseau:inspector (current-window)))

(define-command (com-inspect-syntax :name "Inspect Syntax" :command-table base-table) ()
  (clouseau:inspector (current-syntax)))

-- 
\  Troels
/\ Henriksen


From mathrick at gmail.com  Fri Jun 13 15:04:24 2008
From: mathrick at gmail.com (Maciej Katafiasz)
Date: Fri, 13 Jun 2008 15:04:24 +0000 (UTC)
Subject: [climacs-devel] Build breakage: missing class CHAR-CODE-LIST-LEXEME
Message-ID: <g2u29o$k4i$1@ger.gmane.org>

Hi all,

=====================================================================
There is no class named climacs-prolog-syntax::char-code-list-lexeme.
   [Condition of type simple-error]

Restarts:
 0: [try-recompiling] Recompile prolog2paiprolog and try loading it again
 1: [retry] Retry performing #<asdf:load-op nil {FDD2781}> on #<asdf:cl-
source-file "prolog2paiprolog" {F44FDF9}>.
 2: [accept] Continue, treating #<asdf:load-op nil {FDD2781}> on 
#<asdf:cl-source-file "prolog2paiprolog" {F44FDF9}> as having been 
successful.
 3: [abort] Abort SLIME compilation.
 4: [abort] Return to SLIME's top level.
 5: [terminate-thread] Terminate this thread (#<thread "repl-
thread" {B136C29}>)
=====================================================================

I get that trying to build Climacs (manually) checked out today, along 
with all dependencies (that is, flexistream, McClim, CLX). Any idea what 
it is about? The class references are apparently not new, they have been 
there in a clbuild checkout from long ago which I had, but I don't recall 
it breaking with that.

Cheers,
Maciej




From mathrick at gmail.com  Fri Jun 13 15:27:27 2008
From: mathrick at gmail.com (Maciej Katafiasz)
Date: Fri, 13 Jun 2008 15:27:27 +0000 (UTC)
Subject: [climacs-devel] UI issues: Meta and completion display
Message-ID: <g2u3kv$k4i$2@ger.gmane.org>

Hi,

it seems that there's no way to get minibuffer to display a list of 
possible completions (ie. the thing you get by dashing TAB). That is a 
huge setback if you're trying to see what might be applicable.

Another thing is that Climacs apparently doesn't see my Alt at all, which 
I'd very much like it to see as Meta. Otherwise I'm left with ESC-, and 
that's not the most funniest way to get at things.

Cheers,
Maciej

PS. Related to my blind searching above, is there a REPL to be had in 
Lisp mode?




From athas at sigkill.dk  Fri Jun 13 15:28:12 2008
From: athas at sigkill.dk (Troels Henriksen)
Date: Fri, 13 Jun 2008 17:28:12 +0200
Subject: [climacs-devel] Build breakage: missing class
	CHAR-CODE-LIST-LEXEME
In-Reply-To: <g2u29o$k4i$1@ger.gmane.org> (Maciej Katafiasz's message of "Fri, 
	13 Jun 2008 15:04:24 +0000 (UTC)")
References: <g2u29o$k4i$1@ger.gmane.org>
Message-ID: <871w319wxf.fsf@lambda.athas.dyndns.dk>

Maciej Katafiasz <mathrick at gmail.com> writes:

> Hi all,
>
> =====================================================================
> There is no class named climacs-prolog-syntax::char-code-list-lexeme.
>    [Condition of type simple-error]

Do you mess around with your read-table? In prolog-syntax.lisp:101 is
the expression that creates the name of these lexeme classes. If it's
broken, I would appreciate it if you could figure out how to fix it
(it seems okay to me, though).

-- 
\  Troels
/\ Henriksen


From athas at sigkill.dk  Fri Jun 13 15:33:15 2008
From: athas at sigkill.dk (Troels Henriksen)
Date: Fri, 13 Jun 2008 17:33:15 +0200
Subject: [climacs-devel] UI issues: Meta and completion display
In-Reply-To: <g2u3kv$k4i$2@ger.gmane.org> (Maciej Katafiasz's message of "Fri, 
	13 Jun 2008 15:27:27 +0000 (UTC)")
References: <g2u3kv$k4i$2@ger.gmane.org>
Message-ID: <87tzfx8i4k.fsf@lambda.athas.dyndns.dk>

Maciej Katafiasz <mathrick at gmail.com> writes:

> it seems that there's no way to get minibuffer to display a list of 
> possible completions (ie. the thing you get by dashing TAB). That is a 
> huge setback if you're trying to see what might be applicable.

You can right-click the mouse to get a list of completions, and
there's also some keyboard gesture that I can never remember... try
C-/.

> Another thing is that Climacs apparently doesn't see my Alt at all, which 
> I'd very much like it to see as Meta. Otherwise I'm left with ESC-, and 
> that's not the most funniest way to get at things.

How does xev see your Alt key? Even better, how does CLIM see it? Open
a Lisp syntax buffer in Climacs, evaluate (describe
(clim:read-gesture)), then press your Alt key.

> PS. Related to my blind searching above, is there a REPL to be had in 
> Lisp mode?

No, but if you evaluate stuff from the buffer, output will be visible
in a special view. Also, if you push stuff onto the kill ring from
Climacs, you can yank it in the CLIM Listener.

-- 
\  Troels
/\ Henriksen


From mathrick at gmail.com  Fri Jun 13 15:50:22 2008
From: mathrick at gmail.com (Maciej Katafiasz)
Date: Fri, 13 Jun 2008 15:50:22 +0000 (UTC)
Subject: [climacs-devel] Re: Build breakage: missing
	class	CHAR-CODE-LIST-LEXEME
References: <g2u29o$k4i$1@ger.gmane.org>
	<871w319wxf.fsf@lambda.athas.dyndns.dk>
Message-ID: <g2u4vu$est$1@ger.gmane.org>

Den Fri, 13 Jun 2008 17:28:12 +0200 skrev Troels Henriksen:

> Maciej Katafiasz <mathrick at gmail.com> writes:
> 
>> Hi all,
>>
>> =====================================================================
>> There is no class named climacs-prolog-syntax::char-code-list-lexeme.
>>    [Condition of type simple-error]
> 
> Do you mess around with your read-table? In prolog-syntax.lisp:101 is
> the expression that creates the name of these lexeme classes. If it's
> broken, I would appreciate it if you could figure out how to fix it (it
> seems okay to me, though).

No longer READ-TABLE, but I do mess with *PRINT-CASE* (I have it 
at :DOWNCASE). Lemme take a look at it, though I suspect it's due to 
there being (format nil "~A" 'foo) instead of (format nil "~A" (symbol-
name 'foo)).

Cheers,
Maciej




From mathrick at gmail.com  Fri Jun 13 19:19:03 2008
From: mathrick at gmail.com (Maciej Katafiasz)
Date: Fri, 13 Jun 2008 19:19:03 +0000 (UTC)
Subject: [climacs-devel] Re: Build breakage:
	missing	class	CHAR-CODE-LIST-LEXEME
References: <g2u29o$k4i$1@ger.gmane.org>
	<871w319wxf.fsf@lambda.athas.dyndns.dk> <g2u4vu$est$1@ger.gmane.org>
Message-ID: <g2uh77$u0v$1@ger.gmane.org>

Den Fri, 13 Jun 2008 15:50:22 +0000 skrev Maciej Katafiasz:

> No longer READ-TABLE, but I do mess with *PRINT-CASE* (I have it at
> :DOWNCASE). Lemme take a look at it, though I suspect it's due to there
> being (format nil "~A" 'foo) instead of (format nil "~A" (symbol- name
> 'foo)).

Okay, that was indeed it. Below is a patch, with a couple of remarks:

1) I don't know where to put FORMAT-SYM. It should be in some widely-
accessible place, ideally a CL-UTILS package of sorts, so that hopefully 
people will finally get it right wrt. symbols and case... Anyway, 
Clim(acs) package structure confuses me to no end, what with :CLIM and 
friends not using :CL for example. So here you have the code, and you 
pick where to stick it in.

(defun format-sym (format-string &rest args)
  "Return FORMAT-STRING with args spliced in, where all arguments that
  are symbols with have their SYMBOL-NAME spliced instead, this makes
  sure the result is correct even on systems where read/print case is
  other than default."
  (apply #'format
         nil
         format-string
         (mapcar (lambda (arg)
                   (if (symbolp arg)
                       (symbol-name arg)
                       arg))
                 args)))

2) Not directly related, but I stumbled upon the following odd snippet. 
Why doesn't it simply return "C"?

[c-syntax.lisp:49]

(defmethod name-for-info-pane ((syntax c-syntax) &key pane)
  (declare (ignore pane))
  (format nil "C"))

3) And finally, the patch against Climacs CVS. I tried to identify all 
spots where I believe symbols were being manipulated, tell me if I got 
something wrong or missed anything. Similar treatment would be in order 
for the entirety of McCLIM, I imagine, but I want to stick just to 
Climacs first.

Cheers,
Maciej



cvs diff: Diffing .
Index: climacs-lisp-syntax.lisp
===================================================================
RCS file: /project/climacs/cvsroot/climacs/climacs-lisp-syntax.lisp,v
retrieving revision 1.16
diff -u -r1.16 climacs-lisp-syntax.lisp
--- climacs-lisp-syntax.lisp	5 Feb 2008 22:07:30 -0000	1.16
+++ climacs-lisp-syntax.lisp	13 Jun 2008 19:16:07 -0000
@@ -422,7 +422,7 @@
                   (with-drawing-options (stream :ink +dark-blue+
                                                 :text-style (make-text-
style :fixed nil nil))
                     (princ (dspec item) stream))))
-           (climacs-gui:with-typeout-view (stream (format nil "~A ~A" 
type symbol))
+           (climacs-gui:with-typeout-view (stream (format-sym "~A ~A" 
type symbol))
              (loop for xref in xrefs
                 do (with-output-as-presentation (stream xref 'xref)
                      (printer xref stream))
Index: prolog-syntax.lisp
===================================================================
RCS file: /project/climacs/cvsroot/climacs/prolog-syntax.lisp,v
retrieving revision 1.37
diff -u -r1.37 prolog-syntax.lisp
--- prolog-syntax.lisp	11 Feb 2008 22:50:27 -0000	1.37
+++ prolog-syntax.lisp	13 Jun 2008 19:16:08 -0000
@@ -98,7 +98,7 @@
 (defmethod syntactic-lexeme ((lexeme prolog-lexeme))
   lexeme)
 (macrolet ((def ((name &optional tokenp) &rest subs)
-	     (flet ((f (x) (intern (format nil "~A-~A" x '#:lexeme))))
+	     (flet ((f (x) (intern (format-sym "~A-~A" x '#:lexeme))))
 	       `(progn
 		  (defclass ,(f name) (prolog-lexeme) ())
 
cvs diff: Diffing Doc
cvs diff: Diffing Persistent
cvs diff: Diffing cl-automaton
cvs diff: Diffing testing




From mathrick at gmail.com  Fri Jun 13 19:28:56 2008
From: mathrick at gmail.com (Maciej Katafiasz)
Date: Fri, 13 Jun 2008 19:28:56 +0000 (UTC)
Subject: [climacs-devel] Re: UI issues: Meta and completion display
References: <g2u3kv$k4i$2@ger.gmane.org>
	<87tzfx8i4k.fsf@lambda.athas.dyndns.dk>
Message-ID: <g2uhpo$u2k$1@ger.gmane.org>

Den Fri, 13 Jun 2008 17:33:15 +0200 skrev Troels Henriksen:

> Maciej Katafiasz <mathrick at gmail.com> writes:
> 
>> it seems that there's no way to get minibuffer to display a list of
>> possible completions (ie. the thing you get by dashing TAB). That is a
>> huge setback if you're trying to see what might be applicable.
> 
> You can right-click the mouse to get a list of completions, and there's
> also some keyboard gesture that I can never remember... try C-/.
> 
>> Another thing is that Climacs apparently doesn't see my Alt at all,
>> which I'd very much like it to see as Meta. Otherwise I'm left with
>> ESC-, and that's not the most funniest way to get at things.
> 
> How does xev see your Alt key? Even better, how does CLIM see it? Open a
> Lisp syntax buffer in Climacs, evaluate (describe (clim:read-gesture)),
> then press your Alt key.

Okay, so it says key-name :alt-left, there's more info, but no way to 
copy or paste it, so the best I could do is to give you a screenshot :). 
Incidentally, there's also no way to paste into Climacs from the outside 
world, neither PRIMARY (ie. middle click) nor CLIPBOARD (ie. clipboard) 
seem to work.
 
>> PS. Related to my blind searching above, is there a REPL to be had in
>> Lisp mode?
> 
> No, but if you evaluate stuff from the buffer, output will be visible in
> a special view. Also, if you push stuff onto the kill ring from Climacs,
> you can yank it in the CLIM Listener.

Ah, I see. And is there a way to get a listener inside Climacs?

Cheers,
Maciej



From athas at sigkill.dk  Sun Jun 15 08:57:17 2008
From: athas at sigkill.dk (Troels Henriksen)
Date: Sun, 15 Jun 2008 10:57:17 +0200
Subject: [climacs-devel] Re: UI issues: Meta and completion display
In-Reply-To: <g2uhpo$u2k$1@ger.gmane.org> (Maciej Katafiasz's message of "Fri, 
	13 Jun 2008 19:28:56 +0000 (UTC)")
References: <g2u3kv$k4i$2@ger.gmane.org>
	<87tzfx8i4k.fsf@lambda.athas.dyndns.dk> <g2uhpo$u2k$1@ger.gmane.org>
Message-ID: <874p7vvzwy.fsf@lambda.athas.dyndns.dk>

Maciej Katafiasz <mathrick at gmail.com> writes:

> Okay, so it says key-name :alt-left, there's more info, but no way to 
> copy or paste it, so the best I could do is to give you a screenshot :). 
> Incidentally, there's also no way to paste into Climacs from the outside 
> world, neither PRIMARY (ie. middle click) nor CLIPBOARD (ie. clipboard) 
> seem to work.

Actually, for arcane reasons, you can copy from the *standard-output*
view by holding down Shift and selecting the text with the mouse while
holding down the primary button.

You're correct that there's no way to paste from the outside world
into Climacs, however, and fixing that will require some additional
work on McCLIM (CLIM does not specify a standard way to access the
clipboard).

> Ah, I see. And is there a way to get a listener inside Climacs?

No.

-- 
\  Troels
/\ Henriksen


From athas at sigkill.dk  Sun Jun 15 09:13:07 2008
From: athas at sigkill.dk (Troels Henriksen)
Date: Sun, 15 Jun 2008 11:13:07 +0200
Subject: [climacs-devel] Re: Build breakage:
	missing	class	CHAR-CODE-LIST-LEXEME
In-Reply-To: <g2uh77$u0v$1@ger.gmane.org> (Maciej Katafiasz's message of "Fri, 
	13 Jun 2008 19:19:03 +0000 (UTC)")
References: <g2u29o$k4i$1@ger.gmane.org>
	<871w319wxf.fsf@lambda.athas.dyndns.dk> <g2u4vu$est$1@ger.gmane.org>
	<g2uh77$u0v$1@ger.gmane.org>
Message-ID: <87y757ukm4.fsf@lambda.athas.dyndns.dk>

Maciej Katafiasz <mathrick at gmail.com> writes:

> 1) I don't know where to put FORMAT-SYM. It should be in some widely-
> accessible place, ideally a CL-UTILS package of sorts, so that hopefully 
> people will finally get it right wrt. symbols and case... Anyway, 
> Clim(acs) package structure confuses me to no end, what with :CLIM and 
> friends not using :CL for example. So here you have the code, and you 
> pick where to stick it in.

I put it in ESA-UTILS, which just happens to be the util package I
have available without adding more dependencies.

The package structure for CLIM programs is actually fairy simple:
instead of USEing CL, you USE CLIM-LISP. That's all there is to it.

> 2) Not directly related, but I stumbled upon the following odd snippet. 
> Why doesn't it simply return "C"?
>
> [c-syntax.lisp:49]
>
> (defmethod name-for-info-pane ((syntax c-syntax) &key pane)
>   (declare (ignore pane))
>   (format nil "C"))

Probably just copied from Lisp syntax, which used to create a more
interesting string.

> 3) And finally, the patch against Climacs CVS. I tried to identify all 
> spots where I believe symbols were being manipulated, tell me if I got 
> something wrong or missed anything. Similar treatment would be in order 
> for the entirety of McCLIM, I imagine, but I want to stick just to 
> Climacs first.
Looks fine to me, I just committed it. Thanks!

-- 
\  Troels
/\ Henriksen