matlab-emacs-discuss — Discuss Emacs/MATLAB integration

Re: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

[David C. <dav...@fr...>]

Brilliant, Eric !! Many thanks for your help.

I think I'm making good progress in configuring matlab-emacs and getting
used to it :-)

 

A few more questions, if I may, because I'm really not a GUD expert.
EmacsLink (and most debuggers) provides

 

-          "matlab-eei-step-out" to finish the execution of the current
function, and stop. My understanding is that "gud-finish" should do the
same, but for me it just invokes "dbquit" and stops there. 

 

-          "matlab-eei-go-until-cursor" but although there seems to be a
"gud-until" in gud.el, emacs just doesn't know this function at all (M-x
gud-until can't be invoked). Any clue what the problem is ?

 

-          "matlab-eei-exit-debug" to stop the current session. My
understanding is that "gud-stop-subjob" should do the same, but I only get
the following string echoed in the matlab-shell window: "M-x g u d - s t o p
- s u b j o b RET". 

 

At last, I think I have identified a subtle problem with hyperlinks.

If I start debugging a script, and go step by step, the matlab-shell window
properly shows hyperlinks with the line number, followed by the line of
code. That works fine.

Now say I click on any of the above hyperlinks (without quitting the
debugging session), it brings me back to the source code as expected.

But if I start stepping through my code again, then the hyperlinks are not
parsed correctly any more, and I get something like:

 

<a href="matlab: opentoline('~/matlab/peri.m',106,1)">106 </a>end

 

That's a bit cosmetic really, but it looks like following the hyperlink in
the first place broke something.

 

Thanks in advance for any help you can provide on the GUD functions !

 

Cheers,

David.

 

  _____  

From: Eric Ludlam [mailto:Eri...@ma...] 
Sent: 09 December 2011 13:49
To: David Chappaz; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

 

Hi,

 

for (1), there is a function called : matlab-with-emacs-link you can use.

 

(if (matlab-with-emacs-link)

     (define-key .)

  (define-key.)

 

for (2), that's the way it has always worked.  The keybindings, however,
work in any MATLAB buffer as long as matlab-shell-enable-gud-flag is set,
which is t by default.

 

for (3), I think I've seen that reported, but not seen it myself.  Your
repro steps, however, were good so now I see the problem.

 

On line 4925 the variable 'limitpos' is being set to something based on
characters in the line, including the period.  That means if you complete:

 

mystruct.plotmat <TAB>

 

it will complete ok without saying

 

mystruct.plotmatmystruct.plotmatrix

 

Unfortunately, when you complete for edit, the return completed file name
has a . in it.

 

To fix this, we'd have to add a little hack to know it was completing for
'edit' specifically, and to ignore the .

 

I suppose an alternative would be to take the completion string, and search
for the matching substring in the replacement string, and replace only that.
I don't think there is a utility in Emacs that matches like that.

 

(4) - the figure focus thing is annoying.  The problem here is just MATLAB
doesn't know where the command line is, so can't give focus back.  I don't
know how to fix that.

 

(5) - matlab-shell-run-region runs the region, but gud-print I think picks
some symbol under point to evaluate.   I use run region to highlight a bunch
of lines to run.

 

Eric

Re: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

[Eric L. <Eri...@ma...>]

Hi,

for (1), there is a function called : matlab-with-emacs-link you can use.

(if (matlab-with-emacs-link)
     (define-key ...)
  (define-key...)

for (2), that's the way it has always worked.  The keybindings, however, work in any MATLAB buffer as long as matlab-shell-enable-gud-flag is set, which is t by default.

for (3), I think I've seen that reported, but not seen it myself.  Your repro steps, however, were good so now I see the problem.

On line 4925 the variable 'limitpos' is being set to something based on characters in the line, including the period.  That means if you complete:

mystruct.plotmat <TAB>

it will complete ok without saying

mystruct.plotmatmystruct.plotmatrix

Unfortunately, when you complete for edit, the return completed file name has a . in it.

To fix this, we'd have to add a little hack to know it was completing for 'edit' specifically, and to ignore the .

I suppose an alternative would be to take the completion string, and search for the matching substring in the replacement string, and replace only that.  I don't think there is a utility in Emacs that matches like that.

(4) - the figure focus thing is annoying.  The problem here is just MATLAB doesn't know where the command line is, so can't give focus back.  I don't know how to fix that.

(5) - matlab-shell-run-region runs the region, but gud-print I think picks some symbol under point to evaluate.   I use run region to highlight a bunch of lines to run.

Eric

From: David Chappaz [mailto:dav...@fr...]
Sent: Friday, December 09, 2011 5:08 AM
To: Eric Ludlam; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi Eric,

You should really be thanked for all the hard work you made, and for improving matlab-emacs in the last couple of years since EmacsLink has disappeared.
I happen to be working next door to a Mathworks development site, and I regularly meet some of the Mathworks folks... They always try hard to convince me to move the new Matlab releases, but for now I'm sticking with R2008b because I think EmacsLink is invaluable. I do have R2011b but just don't use it....

Anyway, many thanks for your feedback and possible directions to have "visual breakpoints". It does sound tricky, and way beyond my very limited elisp skills. I'd be prepared to spend some time looking at it, if I thought it was within my abilities, but I'm afraid it's not.

I can't deny I am reluctant to give up EmacsLink, but I'm willing to give matlab-emacs a good go. For now I have a few minor issues which hopefully the community can help me solve:

1/ Because I mostly use EmacsLink, I have configured "Visual Sudio like" key bindings for debugging, e.g.:

(define-key matlab-mode-map [f9]        'matlab-eei-breakpoint-set-clear)
(define-key matlab-mode-map [f10]       'matlab-eei-step)
(define-key matlab-mode-map [f11]       'matlab-eei-step-in)

But now with matlab-emacs, I would have to redefine the bindings to something like:

(define-key matlab-mode-map [f9]        'gud-break)
(define-key matlab-mode-map [f10]       'gud-step)
(define-key matlab-mode-map [f11]       'gud-next)

Ideally I would like my key bindings to work both with EmacsLink and matlab-emacs. Is there a simple way to detect which "mode" is currently used, e.g. something like:

(define-key matlab-mode-map [f9]        '(lambda () (interactive) (if (EmacsLink) (matlab-eei-breakpoint-set-clear) (gud-break))))
(define-key matlab-mode-map [f10]       '(lambda () (interactive) (if (EmacsLink) (matlab-eei-step) (gud-step))))
(define-key matlab-mode-map [f11]       '(lambda () (interactive) (if (EmacsLink) (matlab-eei-step-in) (gud-next))))


2/ The GUD toolbar doesn't automatically show up in my m source files, unless I manually set a breakpoint (e.g. with M-x gud-break), and effectively *** execute *** the function.
The same must be repeated for ALL m files, one by one.
Is this the way it is expected to be ? Or should the GUD toolbar show up automatically in all opened m files ?

3/ I also have a problem with auto-complete.
Say I type "edit plot" and at this point hit TAB. As expected I get the list of all possible completions...  Now say I want to edit "plot.m", so I further type "." and hit TAB again
Then auto-complete gives me the following:

edit plot.plot.m

Can anyone reproduce this ? Or is there something wrong with my config ?

4/ Something which I find really (I mean really) annoying with matlab-emacs is that figures do steal focus.
At the matlab shell prompt, I can't just type "figure", hit ENTER, and continue typing... I have to take my mouse, and return the focus to the matlab shell.... Any workaround for this problem ?

5/ This is more a question than a problem really, but can anyone explain the difference (if any) between matlab-shell-run-region (as in the matlab menu) and gud-print-expression (as in the GUD toolbar) ?

Many thanks in advance to anyone who can help !!
Cheers,
David.

________________________________
From: Eric Ludlam [mailto:Eri...@ma...]
Sent: 08 December 2011 15:52
To: David Chappaz; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi,

Good Luck getting your matlab shell working.  It sounds like a tricky environment.

EmacsLink depended on some Java APIs that aren't available any more.  Most of what you may think of as EmacsLink, however, is the same as matlab-emacs except that matlab-emacs on SourceForge has a bunch of misc bug fixes.  The matlab-shell command was an EmacsLink predecessor, and fortunately, still works.

The breakpoint feature was one of the things not available.  I would be possible to fake it out by reading commands that set breakpoints, and using linemark to remember them, but I imagine that would be quite tricky.  It would probably work to have some command for matlab-shell that calls dbstatus, reads the breakpoints, and dumps out a lisp struct that could be read in to populate a linemark database.  I picture a couple pages of code for that.  You'd have to run the command to show the BP though.

If someone would like to give it a try, I'll be happy to give some pointers.
Eric


From: David Chappaz [mailto:dav...@fr...]
Sent: Thursday, December 08, 2011 10:23 AM
To: Eric Ludlam; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi Eric,

Many thanks for your swift reply !!
In fact I've just realised things are more complicated than that... Eventually I did manage to submit matlab as a Sun Grid Engine job, but then, the communication between emacs and matlab is broken because the two processes do not necessarily run on the same machine.... So it looks like the only way is to submit emacs as a job first, and start matlab normally from there (on the same machine !).

On a slightly different note, I've been using EmacsLink for years, but only now am I starting to look at matlab-emacs. Undoubtedly there are a few things I need to learn and to get used to, but one thing that I miss terribly is the ability to "see" where my breakpoints are in the source code, pretty much in the same way as mlint warnings are reported (with linemark.el).

Is this feature not supported ? Or is something broken in my configuration ?

Cheers,
David.


________________________________
From: Eric Ludlam [mailto:Eri...@ma...]
Sent: 08 December 2011 15:07
To: David Chappaz; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi,

  The variable 'matlab-shell-command' needs to be a string that represents a single executable.  In your case, you want it to just be "qrsh".   You would then set 'matlab-shell-command-switches' to be '("-b" "y" "-q" "single" "-V" ...) with all your switches spelled out.  It seems outlook is messing with my 'quote' marks, so you will need to type it in and not copy what I typed here.

Good Luck
Eric

From: David Chappaz [mailto:dav...@fr...]
Sent: Thursday, December 08, 2011 7:51 AM
To: mat...@li...
Subject: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi everyone,

I'm having a simple problem configuring matlab-shell-command.
The thing is I actually never run matlab locally on the machine I am logged on, but matlab is submitted as a job to a large pool of linux machines with the Sun Grid Engine.

In other words, I don't use the "matlab" command at the shell prompt, but instead I use something like:

qrsh -b y -q single -V -cwd -noshell matlab

Now, everything works fine by default, if I set matlab-shell-command to "matlab".
But if I set matlab-shell-command to "qrsh -b y -q single -V -cwd -noshell matlab", a problem occurs in matlab.el (matlab-shell function), and more specifically on this line :

(switch-to-buffer
      (apply 'make-comint matlab-shell-buffer-name matlab-shell-command
                           nil matlab-shell-command-switches))

which returns the error message
Searching for program: no such file or directory, qrsh -b y -q single -V -cwd -noshell matlab

As a workaround, I have tried defining an alias in my .cshrc, something like:
alias sge_matlab 'qrsh -b y -q single -V -cwd -noshell matlab'

and this works to manually start matlab from the shell. But again if I set matlab-shell-command to "sge_matlab", then I get
apply: Searching for program: no such file or directory, sge_matlab

So what is the proper way to get everything to work in my case ?

Many thanks in advance !
Cheers,
David.

Re: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

[David C. <dav...@fr...>]

Hi Eric,

 

You should really be thanked for all the hard work you made, and for
improving matlab-emacs in the last couple of years since EmacsLink has
disappeared.

I happen to be working next door to a Mathworks development site, and I
regularly meet some of the Mathworks folks. They always try hard to convince
me to move the new Matlab releases, but for now I'm sticking with R2008b
because I think EmacsLink is invaluable. I do have R2011b but just don't use
it..

 

Anyway, many thanks for your feedback and possible directions to have
"visual breakpoints". It does sound tricky, and way beyond my very limited
elisp skills. I'd be prepared to spend some time looking at it, if I thought
it was within my abilities, but I'm afraid it's not.

 

I can't deny I am reluctant to give up EmacsLink, but I'm willing to give
matlab-emacs a good go. For now I have a few minor issues which hopefully
the community can help me solve:

 

1/ Because I mostly use EmacsLink, I have configured "Visual Sudio like" key
bindings for debugging, e.g.:

 

(define-key matlab-mode-map [f9]        'matlab-eei-breakpoint-set-clear)

(define-key matlab-mode-map [f10]       'matlab-eei-step)

(define-key matlab-mode-map [f11]       'matlab-eei-step-in)

 

But now with matlab-emacs, I would have to redefine the bindings to
something like:

 

(define-key matlab-mode-map [f9]        'gud-break)

(define-key matlab-mode-map [f10]       'gud-step)

(define-key matlab-mode-map [f11]       'gud-next)

 

Ideally I would like my key bindings to work both with EmacsLink and
matlab-emacs. Is there a simple way to detect which "mode" is currently
used, e.g. something like:

 

(define-key matlab-mode-map [f9]        '(lambda () (interactive) (if
(EmacsLink) (matlab-eei-breakpoint-set-clear) (gud-break)))) 

(define-key matlab-mode-map [f10]       '(lambda () (interactive) (if
(EmacsLink) (matlab-eei-step) (gud-step)))) 

(define-key matlab-mode-map [f11]       '(lambda () (interactive) (if
(EmacsLink) (matlab-eei-step-in) (gud-next)))) 

 

 

2/ The GUD toolbar doesn't automatically show up in my m source files,
unless I manually set a breakpoint (e.g. with M-x gud-break), and
effectively *** execute *** the function.

The same must be repeated for ALL m files, one by one.

Is this the way it is expected to be ? Or should the GUD toolbar show up
automatically in all opened m files ?

 

3/ I also have a problem with auto-complete.

Say I type "edit plot" and at this point hit TAB. As expected I get the list
of all possible completions.  Now say I want to edit "plot.m", so I further
type "." and hit TAB again

Then auto-complete gives me the following:

 

edit plot.plot.m

 

Can anyone reproduce this ? Or is there something wrong with my config ?

 

4/ Something which I find really (I mean really) annoying with matlab-emacs
is that figures do steal focus. 

At the matlab shell prompt, I can't just type "figure", hit ENTER, and
continue typing. I have to take my mouse, and return the focus to the matlab
shell.. Any workaround for this problem ?

 

5/ This is more a question than a problem really, but can anyone explain the
difference (if any) between matlab-shell-run-region (as in the matlab menu)
and gud-print-expression (as in the GUD toolbar) ?

 

Many thanks in advance to anyone who can help !!
Cheers,

David.

 

  _____  

From: Eric Ludlam [mailto:Eri...@ma...] 
Sent: 08 December 2011 15:52
To: David Chappaz; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

 

Hi,

 

Good Luck getting your matlab shell working.  It sounds like a tricky
environment.

 

EmacsLink depended on some Java APIs that aren't available any more.  Most
of what you may think of as EmacsLink, however, is the same as matlab-emacs
except that matlab-emacs on SourceForge has a bunch of misc bug fixes.  The
matlab-shell command was an EmacsLink predecessor, and fortunately, still
works.

 

The breakpoint feature was one of the things not available.  I would be
possible to fake it out by reading commands that set breakpoints, and using
linemark to remember them, but I imagine that would be quite tricky.  It
would probably work to have some command for matlab-shell that calls
dbstatus, reads the breakpoints, and dumps out a lisp struct that could be
read in to populate a linemark database.  I picture a couple pages of code
for that.  You'd have to run the command to show the BP though.

 

If someone would like to give it a try, I'll be happy to give some pointers.

Eric

 

 

From: David Chappaz [mailto:dav...@fr...] 
Sent: Thursday, December 08, 2011 10:23 AM
To: Eric Ludlam; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

 

Hi Eric,

 

Many thanks for your swift reply !!

In fact I've just realised things are more complicated than that. Eventually
I did manage to submit matlab as a Sun Grid Engine job, but then, the
communication between emacs and matlab is broken because the two processes
do not necessarily run on the same machine.. So it looks like the only way
is to submit emacs as a job first, and start matlab normally from there (on
the same machine !).

 

On a slightly different note, I've been using EmacsLink for years, but only
now am I starting to look at matlab-emacs. Undoubtedly there are a few
things I need to learn and to get used to, but one thing that I miss
terribly is the ability to "see" where my breakpoints are in the source
code, pretty much in the same way as mlint warnings are reported (with
linemark.el).

 

Is this feature not supported ? Or is something broken in my configuration ?

 

Cheers,

David.

 

 

  _____  

From: Eric Ludlam [mailto:Eri...@ma...] 
Sent: 08 December 2011 15:07
To: David Chappaz; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

 

Hi,

 

  The variable 'matlab-shell-command' needs to be a string that represents a
single executable.  In your case, you want it to just be "qrsh".   You would
then set 'matlab-shell-command-switches' to be '("-b" "y" "-q" "single" "-V"
.) with all your switches spelled out.  It seems outlook is messing with my
'quote' marks, so you will need to type it in and not copy what I typed
here.

 

Good Luck

Eric

 

From: David Chappaz [mailto:dav...@fr...] 
Sent: Thursday, December 08, 2011 7:51 AM
To: mat...@li...
Subject: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

 

Hi everyone,

 

I'm having a simple problem configuring matlab-shell-command.

The thing is I actually never run matlab locally on the machine I am logged
on, but matlab is submitted as a job to a large pool of linux machines with
the Sun Grid Engine.

 

In other words, I don't use the "matlab" command at the shell prompt, but
instead I use something like:

 

qrsh -b y -q single -V -cwd -noshell matlab

 

Now, everything works fine by default, if I set matlab-shell-command to
"matlab".

But if I set matlab-shell-command to "qrsh -b y -q single -V -cwd -noshell
matlab", a problem occurs in matlab.el (matlab-shell function), and more
specifically on this line :

 

(switch-to-buffer

      (apply 'make-comint matlab-shell-buffer-name matlab-shell-command

                           nil matlab-shell-command-switches))

 

which returns the error message

Searching for program: no such file or directory, qrsh -b y -q single -V
-cwd -noshell matlab

 

As a workaround, I have tried defining an alias in my .cshrc, something
like:

alias sge_matlab 'qrsh -b y -q single -V -cwd -noshell matlab'

 

and this works to manually start matlab from the shell. But again if I set
matlab-shell-command to "sge_matlab", then I get

apply: Searching for program: no such file or directory, sge_matlab

 

So what is the proper way to get everything to work in my case ?

 

Many thanks in advance !

Cheers,

David.

Re: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

[David C. <dav...@fr...>]

Hi Eric,

 

Many thanks for your swift reply !!

In fact I've just realised things are more complicated than that. Eventually
I did manage to submit matlab as a Sun Grid Engine job, but then, the
communication between emacs and matlab is broken because the two processes
do not necessarily run on the same machine.. So it looks like the only way
is to submit emacs as a job first, and start matlab normally from there (on
the same machine !).

 

On a slightly different note, I've been using EmacsLink for years, but only
now am I starting to look at matlab-emacs. Undoubtedly there are a few
things I need to learn and to get used to, but one thing that I miss
terribly is the ability to "see" where my breakpoints are in the source
code, pretty much in the same way as mlint warnings are reported (with
linemark.el).

 

Is this feature not supported ? Or is something broken in my configuration ?

 

Cheers,

David.

 

 

  _____  

From: Eric Ludlam [mailto:Eri...@ma...] 
Sent: 08 December 2011 15:07
To: David Chappaz; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

 

Hi,

 

  The variable 'matlab-shell-command' needs to be a string that represents a
single executable.  In your case, you want it to just be "qrsh".   You would
then set 'matlab-shell-command-switches' to be '("-b" "y" "-q" "single" "-V"
.) with all your switches spelled out.  It seems outlook is messing with my
'quote' marks, so you will need to type it in and not copy what I typed
here.

 

Good Luck

Eric

 

From: David Chappaz [mailto:dav...@fr...] 
Sent: Thursday, December 08, 2011 7:51 AM
To: mat...@li...
Subject: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

 

Hi everyone,

 

I'm having a simple problem configuring matlab-shell-command.

The thing is I actually never run matlab locally on the machine I am logged
on, but matlab is submitted as a job to a large pool of linux machines with
the Sun Grid Engine.

 

In other words, I don't use the "matlab" command at the shell prompt, but
instead I use something like:

 

qrsh -b y -q single -V -cwd -noshell matlab

 

Now, everything works fine by default, if I set matlab-shell-command to
"matlab".

But if I set matlab-shell-command to "qrsh -b y -q single -V -cwd -noshell
matlab", a problem occurs in matlab.el (matlab-shell function), and more
specifically on this line :

 

(switch-to-buffer

      (apply 'make-comint matlab-shell-buffer-name matlab-shell-command

                           nil matlab-shell-command-switches))

 

which returns the error message

Searching for program: no such file or directory, qrsh -b y -q single -V
-cwd -noshell matlab

 

As a workaround, I have tried defining an alias in my .cshrc, something
like:

alias sge_matlab 'qrsh -b y -q single -V -cwd -noshell matlab'

 

and this works to manually start matlab from the shell. But again if I set
matlab-shell-command to "sge_matlab", then I get

apply: Searching for program: no such file or directory, sge_matlab

 

So what is the proper way to get everything to work in my case ?

 

Many thanks in advance !

Cheers,

David.

Re: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

[Eric L. <Eri...@ma...>]

Hi,

Good Luck getting your matlab shell working.  It sounds like a tricky environment.

EmacsLink depended on some Java APIs that aren't available any more.  Most of what you may think of as EmacsLink, however, is the same as matlab-emacs except that matlab-emacs on SourceForge has a bunch of misc bug fixes.  The matlab-shell command was an EmacsLink predecessor, and fortunately, still works.

The breakpoint feature was one of the things not available.  I would be possible to fake it out by reading commands that set breakpoints, and using linemark to remember them, but I imagine that would be quite tricky.  It would probably work to have some command for matlab-shell that calls dbstatus, reads the breakpoints, and dumps out a lisp struct that could be read in to populate a linemark database.  I picture a couple pages of code for that.  You'd have to run the command to show the BP though.

If someone would like to give it a try, I'll be happy to give some pointers.
Eric


From: David Chappaz [mailto:dav...@fr...]
Sent: Thursday, December 08, 2011 10:23 AM
To: Eric Ludlam; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi Eric,

Many thanks for your swift reply !!
In fact I've just realised things are more complicated than that... Eventually I did manage to submit matlab as a Sun Grid Engine job, but then, the communication between emacs and matlab is broken because the two processes do not necessarily run on the same machine.... So it looks like the only way is to submit emacs as a job first, and start matlab normally from there (on the same machine !).

On a slightly different note, I've been using EmacsLink for years, but only now am I starting to look at matlab-emacs. Undoubtedly there are a few things I need to learn and to get used to, but one thing that I miss terribly is the ability to "see" where my breakpoints are in the source code, pretty much in the same way as mlint warnings are reported (with linemark.el).

Is this feature not supported ? Or is something broken in my configuration ?

Cheers,
David.


________________________________
From: Eric Ludlam [mailto:Eri...@ma...]
Sent: 08 December 2011 15:07
To: David Chappaz; mat...@li...
Subject: RE: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi,

  The variable 'matlab-shell-command' needs to be a string that represents a single executable.  In your case, you want it to just be "qrsh".   You would then set 'matlab-shell-command-switches' to be '("-b" "y" "-q" "single" "-V" ...) with all your switches spelled out.  It seems outlook is messing with my 'quote' marks, so you will need to type it in and not copy what I typed here.

Good Luck
Eric

From: David Chappaz [mailto:dav...@fr...]
Sent: Thursday, December 08, 2011 7:51 AM
To: mat...@li...
Subject: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi everyone,

I'm having a simple problem configuring matlab-shell-command.
The thing is I actually never run matlab locally on the machine I am logged on, but matlab is submitted as a job to a large pool of linux machines with the Sun Grid Engine.

In other words, I don't use the "matlab" command at the shell prompt, but instead I use something like:

qrsh -b y -q single -V -cwd -noshell matlab

Now, everything works fine by default, if I set matlab-shell-command to "matlab".
But if I set matlab-shell-command to "qrsh -b y -q single -V -cwd -noshell matlab", a problem occurs in matlab.el (matlab-shell function), and more specifically on this line :

(switch-to-buffer
      (apply 'make-comint matlab-shell-buffer-name matlab-shell-command
                           nil matlab-shell-command-switches))

which returns the error message
Searching for program: no such file or directory, qrsh -b y -q single -V -cwd -noshell matlab

As a workaround, I have tried defining an alias in my .cshrc, something like:
alias sge_matlab 'qrsh -b y -q single -V -cwd -noshell matlab'

and this works to manually start matlab from the shell. But again if I set matlab-shell-command to "sge_matlab", then I get
apply: Searching for program: no such file or directory, sge_matlab

So what is the proper way to get everything to work in my case ?

Many thanks in advance !
Cheers,
David.

Re: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

[Eric L. <Eri...@ma...>]

Hi,

  The variable 'matlab-shell-command' needs to be a string that represents a single executable.  In your case, you want it to just be "qrsh".   You would then set 'matlab-shell-command-switches' to be '("-b" "y" "-q" "single" "-V" ...) with all your switches spelled out.  It seems outlook is messing with my 'quote' marks, so you will need to type it in and not copy what I typed here.

Good Luck
Eric

From: David Chappaz [mailto:dav...@fr...]
Sent: Thursday, December 08, 2011 7:51 AM
To: mat...@li...
Subject: [Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

Hi everyone,

I'm having a simple problem configuring matlab-shell-command.
The thing is I actually never run matlab locally on the machine I am logged on, but matlab is submitted as a job to a large pool of linux machines with the Sun Grid Engine.

In other words, I don't use the "matlab" command at the shell prompt, but instead I use something like:

qrsh -b y -q single -V -cwd -noshell matlab

Now, everything works fine by default, if I set matlab-shell-command to "matlab".
But if I set matlab-shell-command to "qrsh -b y -q single -V -cwd -noshell matlab", a problem occurs in matlab.el (matlab-shell function), and more specifically on this line :

(switch-to-buffer
      (apply 'make-comint matlab-shell-buffer-name matlab-shell-command
                           nil matlab-shell-command-switches))

which returns the error message
Searching for program: no such file or directory, qrsh -b y -q single -V -cwd -noshell matlab

As a workaround, I have tried defining an alias in my .cshrc, something like:
alias sge_matlab 'qrsh -b y -q single -V -cwd -noshell matlab'

and this works to manually start matlab from the shell. But again if I set matlab-shell-command to "sge_matlab", then I get
apply: Searching for program: no such file or directory, sge_matlab

So what is the proper way to get everything to work in my case ?

Many thanks in advance !
Cheers,
David.

[Matlab-emacs-discuss] Matlab-Shell & Sun Grid Engine

[David C. <dav...@fr...>]

Hi everyone,

 

I'm having a simple problem configuring matlab-shell-command.

The thing is I actually never run matlab locally on the machine I am logged
on, but matlab is submitted as a job to a large pool of linux machines with
the Sun Grid Engine.

 

In other words, I don't use the "matlab" command at the shell prompt, but
instead I use something like:

 

qrsh -b y -q single -V -cwd -noshell matlab

 

Now, everything works fine by default, if I set matlab-shell-command to
"matlab".

But if I set matlab-shell-command to "qrsh -b y -q single -V -cwd -noshell
matlab", a problem occurs in matlab.el (matlab-shell function), and more
specifically on this line :

 

(switch-to-buffer

      (apply 'make-comint matlab-shell-buffer-name matlab-shell-command

                           nil matlab-shell-command-switches))

 

which returns the error message

Searching for program: no such file or directory, qrsh -b y -q single -V
-cwd -noshell matlab

 

As a workaround, I have tried defining an alias in my .cshrc, something
like:

alias sge_matlab 'qrsh -b y -q single -V -cwd -noshell matlab'

 

and this works to manually start matlab from the shell. But again if I set
matlab-shell-command to "sge_matlab", then I get

apply: Searching for program: no such file or directory, sge_matlab

 

So what is the proper way to get everything to work in my case ?

 

Many thanks in advance !

Cheers,

David.

[Matlab-emacs-discuss] Problem defining include paths?

[Matthieu L. <mat...@ho...>]

Hello,

I am trying to get semantic to work with my Matlab projects. I have the following defined in semanticdb-matlab.el:

(defvar semanticdb-matlab-include-paths
      '("/home1/m/matthieu/projects/"))


When I open a Matlab file under my "projects" in emacs, semanticdb spends some time "Searching for user files" (about 30 seconds). Then I select a function call that is defined somewhere else under "projects" and execute "semantic-ia-fast-jump". It fails with the message "Could not find suitable jump point for [my function name]". However, *after* opening the file where the function is defined, the jump works.

I guess the semanticdb indexing function is not working for me, but I don't know what I am missing. Does the indexing work recursively? Is semanticdb-matlab-include-paths the right variable to use? (I recompiled all emacs-matlab files after adding my project directory and it still doesn't work).

Thank you for your help,
Matthieu
 		 	   		  

Re: [Matlab-emacs-discuss] Bug report (patch): matlab.el: previous command at shell prompt fix

[Mark H. <mh...@gm...>]

Thanks for the reply.

I've been testing this over the past few weeks, and the issue still pops up sporadically for me, so you should not include the patch below.

I think it may be related to using Aquamacs on Mac OS X.  
I've switched to using a more stock Emacs: 23.3 from http://emacsformacosx.com/builds.
So far I haven't seen it, but if recurs I will look into it and let you know.
Aquamacs seems to have multiple issues with matlab-shell, for anyone else that was thinking of using it.

Best,
Mark


On Oct 5, 2011, at 8:48 AM, Eric Ludlam wrote:

> Hi,
> 
> Sorry for the delay in replying.
> 
> I was not able to reproduce this issue, but the patch seems to continue working fine for me, so I'll include it in the next check-in unless I discover something else along the way.
> 
> Thanks!
> Eric
> 
> -----Original Message-----
> From: Mark Histed [mailto:mh...@gm...] 
> Sent: Thursday, September 08, 2011 1:58 PM
> To: David Engster
> Cc: Eric Ludlam; mat...@li...
> Subject: Bug report (patch): matlab.el: previous command at shell prompt fix
> 
> Hi,
> 
> Using the latest CVS matlab.el, and Matlab 2011a, pressing M-p or the up arrow does not recall previous history.  
> Status line shows "History item 1", but the previous command gets inserted before the ">>" prompt and pressing RET gives an error like "Unknown operator >>" from Matlab.
> 
> This patch fixes it.  But it is unlikely to be the best solution.
> 
> Best
> Mark
> 
> Subject: [PATCH] kludgy fix to make previous command history work at shell prompt
> 
> ---
> matlab.el |    1 +
> 1 files changed, 1 insertions(+), 0 deletions(-)
> 
> diff --git a/matlab.el b/matlab.el
> index fee2bb9..d928458 100644
> --- a/matlab.el
> +++ b/matlab.el
> @@ -4729,6 +4729,7 @@ end\n"
> (defun matlab-shell-previous-matching-input-from-input (n)
>   "Get the Nth previous matching input from for the command line."
>   (interactive "p")
> +  (end-of-line)
>   (if (comint-after-pmark-p)
>       (if (memq last-command '(matlab-shell-previous-matching-input-from-input
>                               matlab-shell-next-matching-input-from-input))
> -- 
> 1.7.0.3
> 
> 

Re: [Matlab-emacs-discuss] Cedet integration

[Eric L. <Eri...@ma...>]

Hi,

I was not able to reproduce this issue (I don't have a mac handy).  Your emacs should have CEDET built in, so perhaps it is related to built-in CEDET not being overridden by cedet 1.0 which you installed separately?  I never tried CEDET from Emacs 23.3 and matlab together, I've always used CEDET installed from CVS or, more recently, BZR.

If you have the CEDET setup from the instructions that came with the cedet 1.0 tar ball you should be ok.

Eric

-----Original Message-----
From: Cláudia Soares [mailto:cla...@gm...] 
Sent: Thursday, September 29, 2011 7:17 PM
To: mat...@li...
Subject: [Matlab-emacs-discuss] Cedet integration

Hi

I'm trying to have a fully functional matlab-emacs + semantic installation, but I'm missing something. When typing code, sometimes I get this error:

incremental parser error: "No support found to parse buffer \"problemInicialization.m\""

This seems to happen whenever I type a struct accessor (a dot) after a struct variable name. Can somebody give me a hint on what might be wrong with my setup?

I'm running MacPorts Emacs 23.3(9.0), cedet 1.0, matlab-mode 3.3.1. In my init file all matlab-emacs related definitions are:


(add-to-list 'load-path (concat elisp-dir "matlab-emacs")) (require 'matlab-load)
(matlab-cedet-setup)
(setq matlab-indent-function-body t)

Thanks
Claudia
------------------------------------------------------------------------------
All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense.
http://p.sf.net/sfu/splunk-d2dcopy1
_______________________________________________
Matlab-emacs-discuss mailing list
Mat...@li...
https://lists.sourceforge.net/lists/listinfo/matlab-emacs-discuss

Re: [Matlab-emacs-discuss] Bug report (patch): matlab.el: previous command at shell prompt fix

[Eric L. <Eri...@ma...>]

Hi,

Sorry for the delay in replying.

I was not able to reproduce this issue, but the patch seems to continue working fine for me, so I'll include it in the next check-in unless I discover something else along the way.

Thanks!
Eric

-----Original Message-----
From: Mark Histed [mailto:mh...@gm...] 
Sent: Thursday, September 08, 2011 1:58 PM
To: David Engster
Cc: Eric Ludlam; mat...@li...
Subject: Bug report (patch): matlab.el: previous command at shell prompt fix

Hi,

Using the latest CVS matlab.el, and Matlab 2011a, pressing M-p or the up arrow does not recall previous history.  
Status line shows "History item 1", but the previous command gets inserted before the ">>" prompt and pressing RET gives an error like "Unknown operator >>" from Matlab.

This patch fixes it.  But it is unlikely to be the best solution.

Best
Mark

Subject: [PATCH] kludgy fix to make previous command history work at shell prompt

---
 matlab.el |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/matlab.el b/matlab.el
index fee2bb9..d928458 100644
--- a/matlab.el
+++ b/matlab.el
@@ -4729,6 +4729,7 @@ end\n"
 (defun matlab-shell-previous-matching-input-from-input (n)
   "Get the Nth previous matching input from for the command line."
   (interactive "p")
+  (end-of-line)
   (if (comint-after-pmark-p)
       (if (memq last-command '(matlab-shell-previous-matching-input-from-input
                               matlab-shell-next-matching-input-from-input))
-- 
1.7.0.3

Re: [Matlab-emacs-discuss] repeated message in matlab-shell: Old-style plist parser

[Timothy W. H. <hi...@un...>]

In case anyone ever runs across this thread trying to solve this
problem, this is a bug in MATLAB R2010b Service Pack 1 for Macs.
There's a simple fix here:

http://www.mathworks.com/support/bugreports/711742

For some reason, that link wasn't easy to find with google...

-Tim

> Hello,
>
> Thanks for the great emacs tool -- I've recently migrated back to Matlab
> after a number of years using mostly other tools and I'm really glad
> this package is available!
>
> I am getting this message repeatedly during matlab-shell sessions in emacs:
>
> 2011-09-23 09:28:18.168 MATLAB_maci64[9988:3003]
>  CFPropertyListCreateFromXMLData(): Old-style plist parser: missing
>  semicolon in dictionary.
>
> From the message text I think the error is coming from Matlab, not from
> matlab-emacs, but it only appears in matlab-shell in emacs.  If I run
> the Matlab GUI I do not see the message.  I posted to this list in hopes
> that someone else might have seen this or know how to fix this?  Or,
> short of fixing it, how to suppress the message?  It's annoying because
> it takes the focus of matlab-shell so that running code regions with C-c
> C-r does not work.
>
> I am running Matlab Version 7.11.1.866 (R2010b) Service Pack 1 64-bit
> (maci64) on OS X 10.5.8.
>
> Thanks,
> Tim
>
> --
>
> Timothy W. Hilton
> Postdoctoral Fellow, Department of Biology
> University of New Mexico
> hilton@...

[Matlab-emacs-discuss] Cedet integration

[Cláudia S. <cla...@gm...>]

Hi

I'm trying to have a fully functional matlab-emacs + semantic installation, but I'm missing something. When typing code, sometimes I get this error:

incremental parser error: "No support found to parse buffer \"problemInicialization.m\""

This seems to happen whenever I type a struct accessor (a dot) after a struct variable name. Can somebody give me a hint on what might be wrong with my setup?

I'm running MacPorts Emacs 23.3(9.0), cedet 1.0, matlab-mode 3.3.1. In my init file all matlab-emacs related definitions are:


(add-to-list 'load-path (concat elisp-dir "matlab-emacs"))
(require 'matlab-load)
(matlab-cedet-setup)
(setq matlab-indent-function-body t)

Thanks
Claudia

Re: [Matlab-emacs-discuss] Fwd: how to see mlint error messages..

[Eric L. <Eri...@ma...>]

Hi,

Sorry for the very slow reply.

Supposedly, if you say 'yes' to configure your mlint, you can point that variable at your mlint programs.  This is a list of places where mlint might be.  If you set it to some places for mlint, it will attempt to find them either on your path, or in a MATLAB install area.  If in the custom buffer you save for a future session, you should be ok.

If it comes back at you a few times, just add mlint to your path before Emacs starts, and that will definitely fix it.

Eric

From: socrates smith [mailto:css...@gm...]
Sent: Tuesday, August 16, 2011 5:52 AM
To: mat...@li...
Subject: [Matlab-emacs-discuss] Fwd: how to see mlint error messages..

Hi,

Every time I enable mlint warnings using,

(setq matlab-show-mlint-warnings t)
(setq matlab-highlight-cross-function-variables t)

emacs says that mlint program has not been configured and asks me to configure it. How do I set a program as the mlint program?

Thanks,
S. Smith

[Matlab-emacs-discuss] repeated message in matlab-shell: Old-style plist parser

[Timothy W. H. <hi...@un...>]

Hello,

Thanks for the great emacs tool -- I've recently migrated back to Matlab
after a number of years using mostly other tools and I'm really glad
this package is available!

I am getting this message repeatedly during matlab-shell sessions in emacs:

2011-09-23 09:28:18.168 MATLAB_maci64[9988:3003]
 CFPropertyListCreateFromXMLData(): Old-style plist parser: missing
 semicolon in dictionary.

>From the message text I think the error is coming from Matlab, not from
matlab-emacs, but it only appears in matlab-shell in emacs.  If I run
the Matlab GUI I do not see the message.  I posted to this list in hopes
that someone else might have seen this or know how to fix this?  Or,
short of fixing it, how to suppress the message?  It's annoying because
it takes the focus of matlab-shell so that running code regions with C-c
C-r does not work.

I am running Matlab Version 7.11.1.866 (R2010b) Service Pack 1 64-bit
(maci64) on OS X 10.5.8.

Thanks,
Tim

--

Timothy W. Hilton
Postdoctoral Fellow, Department of Biology
University of New Mexico
hi...@un...

[Matlab-emacs-discuss] Bug report (patch): matlab.el: previous command at shell prompt fix

[Mark H. <mh...@gm...>]

Hi,

Using the latest CVS matlab.el, and Matlab 2011a, pressing M-p or the up arrow does not recall previous history.  
Status line shows "History item 1", but the previous command gets inserted before the ">>" prompt and pressing RET gives an error like "Unknown operator >>" from Matlab.

This patch fixes it.  But it is unlikely to be the best solution.

Best
Mark

Subject: [PATCH] kludgy fix to make previous command history work at shell prompt

---
 matlab.el |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/matlab.el b/matlab.el
index fee2bb9..d928458 100644
--- a/matlab.el
+++ b/matlab.el
@@ -4729,6 +4729,7 @@ end\n"
 (defun matlab-shell-previous-matching-input-from-input (n)
   "Get the Nth previous matching input from for the command line."
   (interactive "p")
+  (end-of-line)
   (if (comint-after-pmark-p)
       (if (memq last-command '(matlab-shell-previous-matching-input-from-input
                               matlab-shell-next-matching-input-from-input))
-- 
1.7.0.3

[Matlab-emacs-discuss] Fwd: how to see mlint error messages..

[socrates s. <css...@gm...>]

Hi,

Every time I enable mlint warnings using,

(setq matlab-show-mlint-warnings t)
(setq matlab-highlight-cross-function-variables t)

emacs says that mlint program has not been configured and asks me to
configure it. How do I set a program as the mlint program?

Thanks,
S. Smith

Re: [Matlab-emacs-discuss] Insert break-points using matlab-emacs

[socrates s. <css...@gm...>]

Thanks Eric!

On Tue, Jul 26, 2011 at 7:12 PM, Eric Ludlam <Eri...@ma...>wrote:

>  Hi,****
>
> ** **
>
> You can  set breakpoints, and step over code a line at a time.  This is all
> done via GUD interface in Emacs (Grand Unified Debugger).****
>
> ** **
>
> The GUD menu is enabled in matlab files when you hit a break point.  The
> keybinding is ****
>
> ** **
>
> C-x C-a b****
>
> ** **
>
> but if it doesn’t get set, you can always do:****
>
> ** **
>
> M-x gud-break RET****
>
> ** **
>
> on some line of code.****
>
> ** **
>
> In MATLAB Shell, you can do next/step with C-x C-n and C-x C-s.  From the
> source file, the prefix is C-x C-a.****
>
> ** **
>
> This is the standard GUD interface for debuggers.****
>
> ** **
>
> Today I founded my latest Matlab resisting showing the arrow during
> debugging every time.  It may be the process filters in Emacs need to be
> updated with recent changes.****
>
> ** **
>
> Eric****
>
> ** **
>
> *From:* socrates smith [mailto:css...@gm...]
> *Sent:* Wednesday, July 20, 2011 5:20 AM
> *To:* mat...@li...
> *Subject:* [Matlab-emacs-discuss] Insert break-points using matlab-emacs**
> **
>
> ** **
>
> Hi all,
>
> Is it possible to insert breakpoints using the matlab-emacs interface.
> Also, will I be able to run through the code step by step using (F5 or F10)
> or any such key such as in matlab?
>
> Thanks,
> Socrates****
>

Re: [Matlab-emacs-discuss] Possible bug in matlab-shell-run-region when region contains fprintf and sprintf formatting characters

[David E. <de...@ra...>]

Graham Taylor writes:
> Hello,I'm having problems when executing a region with C-c C-r
> (matlab-shell-run-region) when the region contains a line like:
> fprintf('My number is %d\n",4); This also happens with sprintf
> lines. It seems that matlab-emacs interprets everything after the % as
> a comment, and so only the first part of the line is executed.  I'm
> not an emacs lisp hacker at all. Does anyone have a solution?  Graham

That bug should be already fixed; please update from CVS.

-David

[Matlab-emacs-discuss] Possible bug in matlab-shell-run-region when region contains fprintf and sprintf formatting characters

[Graham T. <ess...@ho...>]

Hello,I'm having problems when executing a region with C-c C-r (matlab-shell-run-region) when the region contains a line like:
fprintf('My number is %d\n",4);
This also happens with sprintf lines. It seems that matlab-emacs interprets everything after the % as a comment, and so only the first part of the line is executed.
I'm not an emacs lisp hacker at all. Does anyone have a solution?
Graham 		 	   		   		 	   		  

Re: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?

[Eric L. <Eri...@ma...>]

Hi,

When you start MATLAB in the matlab-shell in Emacs, it causes Matlab to run 'emacsinit'.  In Emacs init, it sets up emacsclient as the default editor.  If your Emacs uses 'server-start' the, typing 'edit' in Matlab will cause the file to open in Emacs.

Eric

-----Original Message-----
From: Laurent Jacques [mailto:lau...@uc...] 
Sent: Friday, July 29, 2011 8:29 AM
To: mat...@li...
Subject: Re: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?

Dear all,

Based on the previous exchange of emails with Eric, I would like now to standardize a bit this ability to redirect fprintf output to a shell.

In particular, I'd like to design the equivalent of a figure in matlab but for text, let's call it a "message box", so that text messages are sent to it.

More clearly, I'd like to achieve something like

%% Opening a new buffer in emacs with a shell running inside  >> open_msg_box

%% get the current message box opened (a label to it)  >> gc_msg_box ans = 'ttys0'

%% print a message from matlab to this shell  >> message("Hello World");


Here is my related question:

Do you know if there is a way (e.g. with emacsclient) to open a new emacs buffer from matlab (running inside emacs in matlab-shell) and to run something inside (like "shell") ?

Best,
Laurent




On 26/07/11 15:13, Laurent Jacques wrote:
> This works great! Thanks Eric.
>
> Under Mac OS, once a terminal is open through emacs or not (say in
> /dev/ttys001)
> I can do now, inside a matlab script,
>
> fid=fopen('/dev/ttys001','w');
> fprintf(fid,'hello world\n');
>
> and this message is truly printed in this terminal!
> This is nothing but the stdout redirection I was looking for.
>
> Best,
> Laurent
>
> On 26/07/11 14:54, Eric Ludlam wrote:
>> I used to do this back in 1996 or so, and it worked like this:
>>
>> Use:
>> M-x shell RET
>>
>> to create a shell, or create a shell with a name by using C-u prefix to the shell command.
>>
>> In the shell, you can find out which terminal it used.  Sometimes it is an environment variable.  I found I can use the 'tty' command on my linux box today.
>>
>> You can fopen that file name and write to it as long as permissions are set ok.
>>
>> I tried this in my emacs today with
>>
>> echo "moose">>   /dev/pts/4
>>
>> and that worked to get to my other shell window.
>>
>> Good Luck
>> Eric
>>
>> -----Original Message-----
>> From: Laurent Jacques [mailto:lau...@uc...]
>> Sent: Tuesday, July 26, 2011 6:41 AM
>> To: mat...@li...
>> Subject: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?
>>
>> Dear all,
>>
>> Question: Do you know if it is possible to redirect the output of a matlab disp()/fprintf() (possibly using another command) to another emacs buffer than the one running matlab-shell ?
>>
>> Objective: When running a "cell" inside a matlab script, this possibility would allow me to display some computation messages or results without to be disturbed by the script code that is always recopied in the matlab shell with the fprintf/disp output.
>>
>>
>> Best regards,
>> Laurent
>>
>>
>>
>>
>> --
>> Dr. Laurent JACQUES
>> Postdoctoral Researcher, Return Grant, BELSPO, Belgium
>>
>> Institute of Information and Communication Technologies,
>> 	Electronics and Applied Mathematics (ICTEAM-TELE) University of 
>> Louvain Batiment Stévin Place du Levant 2, PO Box L5.04.04
>> 1348 Louvain-la-Neuve, Belgium
>>
>> Office: a.133
>> Tel: +32 10 47 81 24
>> Fax: +32 10 47 20 89
>> http://perso.uclouvain.be/laurent.jacques
>>
>>
>> ---------------------------------------------------------------------
>> --------- Magic Quadrant for Content-Aware Data Loss Prevention 
>> Research study explores the data loss prevention market. Includes in-depth analysis on the changes within the DLP market, and the criteria used to evaluate the strengths and weaknesses of these DLP solutions.
>> http://www.accelacomm.com/jaw/sfnl/114/51385063/
>> _______________________________________________
>> Matlab-emacs-discuss mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matlab-emacs-discuss
>


--
Dr. Laurent JACQUES
Postdoctoral Researcher, Return Grant, BELSPO, Belgium

Institute of Information and Communication Technologies,
	Electronics and Applied Mathematics (ICTEAM-TELE) University of Louvain Batiment Stévin Place du Levant 2, PO Box L5.04.04
1348 Louvain-la-Neuve, Belgium

Office: a.133
Tel: +32 10 47 81 24
Fax: +32 10 47 20 89
http://perso.uclouvain.be/laurent.jacques


------------------------------------------------------------------------------
Got Input?   Slashdot Needs You.
Take our quick survey online.  Come on, we don't ask for help often.
Plus, you'll get a chance to win $100 to spend on ThinkGeek.
http://p.sf.net/sfu/slashdot-survey
_______________________________________________
Matlab-emacs-discuss mailing list
Mat...@li...
https://lists.sourceforge.net/lists/listinfo/matlab-emacs-discuss

Re: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?

[Laurent J. <lau...@uc...>]

Dear all,

Based on the previous exchange of emails with Eric, I would like now to 
standardize a bit this ability to redirect fprintf output to a shell.

In particular, I'd like to design the equivalent of a figure in matlab 
but for text, let's call it a "message box", so that text messages are 
sent to it.

More clearly, I'd like to achieve something like

%% Opening a new buffer in emacs with a shell running inside
 >> open_msg_box

%% get the current message box opened (a label to it)
 >> gc_msg_box
ans = 'ttys0'

%% print a message from matlab to this shell
 >> message("Hello World");


Here is my related question:

Do you know if there is a way (e.g. with emacsclient) to open a new 
emacs buffer from matlab (running inside emacs in matlab-shell) and to 
run something inside (like "shell") ?

Best,
Laurent




On 26/07/11 15:13, Laurent Jacques wrote:
> This works great! Thanks Eric.
>
> Under Mac OS, once a terminal is open through emacs or not (say in
> /dev/ttys001)
> I can do now, inside a matlab script,
>
> fid=fopen('/dev/ttys001','w');
> fprintf(fid,'hello world\n');
>
> and this message is truly printed in this terminal!
> This is nothing but the stdout redirection I was looking for.
>
> Best,
> Laurent
>
> On 26/07/11 14:54, Eric Ludlam wrote:
>> I used to do this back in 1996 or so, and it worked like this:
>>
>> Use:
>> M-x shell RET
>>
>> to create a shell, or create a shell with a name by using C-u prefix to the shell command.
>>
>> In the shell, you can find out which terminal it used.  Sometimes it is an environment variable.  I found I can use the 'tty' command on my linux box today.
>>
>> You can fopen that file name and write to it as long as permissions are set ok.
>>
>> I tried this in my emacs today with
>>
>> echo "moose">>   /dev/pts/4
>>
>> and that worked to get to my other shell window.
>>
>> Good Luck
>> Eric
>>
>> -----Original Message-----
>> From: Laurent Jacques [mailto:lau...@uc...]
>> Sent: Tuesday, July 26, 2011 6:41 AM
>> To: mat...@li...
>> Subject: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?
>>
>> Dear all,
>>
>> Question: Do you know if it is possible to redirect the output of a matlab disp()/fprintf() (possibly using another command) to another emacs buffer than the one running matlab-shell ?
>>
>> Objective: When running a "cell" inside a matlab script, this possibility would allow me to display some computation messages or results without to be disturbed by the script code that is always recopied in the matlab shell with the fprintf/disp output.
>>
>>
>> Best regards,
>> Laurent
>>
>>
>>
>>
>> --
>> Dr. Laurent JACQUES
>> Postdoctoral Researcher, Return Grant, BELSPO, Belgium
>>
>> Institute of Information and Communication Technologies,
>> 	Electronics and Applied Mathematics (ICTEAM-TELE) University of Louvain Batiment Stévin Place du Levant 2, PO Box L5.04.04
>> 1348 Louvain-la-Neuve, Belgium
>>
>> Office: a.133
>> Tel: +32 10 47 81 24
>> Fax: +32 10 47 20 89
>> http://perso.uclouvain.be/laurent.jacques
>>
>>
>> ------------------------------------------------------------------------------
>> Magic Quadrant for Content-Aware Data Loss Prevention Research study explores the data loss prevention market. Includes in-depth analysis on the changes within the DLP market, and the criteria used to evaluate the strengths and weaknesses of these DLP solutions.
>> http://www.accelacomm.com/jaw/sfnl/114/51385063/
>> _______________________________________________
>> Matlab-emacs-discuss mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matlab-emacs-discuss
>


-- 
Dr. Laurent JACQUES
Postdoctoral Researcher, Return Grant, BELSPO, Belgium

Institute of Information and Communication Technologies,
	Electronics and Applied Mathematics (ICTEAM-TELE)
University of Louvain
Batiment Stévin
Place du Levant 2, PO Box L5.04.04
1348 Louvain-la-Neuve, Belgium

Office: a.133
Tel: +32 10 47 81 24
Fax: +32 10 47 20 89
http://perso.uclouvain.be/laurent.jacques

Re: [Matlab-emacs-discuss] Insert break-points using matlab-emacs

[Eric L. <Eri...@ma...>]

Hi,

You can  set breakpoints, and step over code a line at a time.  This is all done via GUD interface in Emacs (Grand Unified Debugger).

The GUD menu is enabled in matlab files when you hit a break point.  The keybinding is

C-x C-a b

but if it doesn't get set, you can always do:

M-x gud-break RET

on some line of code.

In MATLAB Shell, you can do next/step with C-x C-n and C-x C-s.  From the source file, the prefix is C-x C-a.

This is the standard GUD interface for debuggers.

Today I founded my latest Matlab resisting showing the arrow during debugging every time.  It may be the process filters in Emacs need to be updated with recent changes.

Eric

From: socrates smith [mailto:css...@gm...]
Sent: Wednesday, July 20, 2011 5:20 AM
To: mat...@li...
Subject: [Matlab-emacs-discuss] Insert break-points using matlab-emacs

Hi all,

Is it possible to insert breakpoints using the matlab-emacs interface. Also, will I be able to run through the code step by step using (F5 or F10) or any such key such as in matlab?

Thanks,
Socrates

Re: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?

[Laurent J. <lau...@uc...>]

This works great! Thanks Eric.

Under Mac OS, once a terminal is open through emacs or not (say in 
/dev/ttys001)
I can do now, inside a matlab script,

fid=fopen('/dev/ttys001','w');
fprintf(fid,'hello world\n');

and this message is truly printed in this terminal!
This is nothing but the stdout redirection I was looking for.

Best,
Laurent

On 26/07/11 14:54, Eric Ludlam wrote:
> I used to do this back in 1996 or so, and it worked like this:
>
> Use:
> M-x shell RET
>
> to create a shell, or create a shell with a name by using C-u prefix to the shell command.
>
> In the shell, you can find out which terminal it used.  Sometimes it is an environment variable.  I found I can use the 'tty' command on my linux box today.
>
> You can fopen that file name and write to it as long as permissions are set ok.
>
> I tried this in my emacs today with
>
> echo "moose">>  /dev/pts/4
>
> and that worked to get to my other shell window.
>
> Good Luck
> Eric
>
> -----Original Message-----
> From: Laurent Jacques [mailto:lau...@uc...]
> Sent: Tuesday, July 26, 2011 6:41 AM
> To: mat...@li...
> Subject: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?
>
> Dear all,
>
> Question: Do you know if it is possible to redirect the output of a matlab disp()/fprintf() (possibly using another command) to another emacs buffer than the one running matlab-shell ?
>
> Objective: When running a "cell" inside a matlab script, this possibility would allow me to display some computation messages or results without to be disturbed by the script code that is always recopied in the matlab shell with the fprintf/disp output.
>
>
> Best regards,
> Laurent
>
>
>
>
> --
> Dr. Laurent JACQUES
> Postdoctoral Researcher, Return Grant, BELSPO, Belgium
>
> Institute of Information and Communication Technologies,
> 	Electronics and Applied Mathematics (ICTEAM-TELE) University of Louvain Batiment Stévin Place du Levant 2, PO Box L5.04.04
> 1348 Louvain-la-Neuve, Belgium
>
> Office: a.133
> Tel: +32 10 47 81 24
> Fax: +32 10 47 20 89
> http://perso.uclouvain.be/laurent.jacques
>
>
> ------------------------------------------------------------------------------
> Magic Quadrant for Content-Aware Data Loss Prevention Research study explores the data loss prevention market. Includes in-depth analysis on the changes within the DLP market, and the criteria used to evaluate the strengths and weaknesses of these DLP solutions.
> http://www.accelacomm.com/jaw/sfnl/114/51385063/
> _______________________________________________
> Matlab-emacs-discuss mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matlab-emacs-discuss


-- 
Dr. Laurent JACQUES
Postdoctoral Researcher, Return Grant, BELSPO, Belgium

Institute of Information and Communication Technologies,
	Electronics and Applied Mathematics (ICTEAM-TELE)
University of Louvain
Batiment Stévin
Place du Levant 2, PO Box L5.04.04
1348 Louvain-la-Neuve, Belgium

Office: a.133
Tel: +32 10 47 81 24
Fax: +32 10 47 20 89
http://perso.uclouvain.be/laurent.jacques

Re: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?

[Eric L. <Eri...@ma...>]

I used to do this back in 1996 or so, and it worked like this:

Use:
M-x shell RET

to create a shell, or create a shell with a name by using C-u prefix to the shell command.

In the shell, you can find out which terminal it used.  Sometimes it is an environment variable.  I found I can use the 'tty' command on my linux box today.

You can fopen that file name and write to it as long as permissions are set ok.

I tried this in my emacs today with

echo "moose" >> /dev/pts/4

and that worked to get to my other shell window.

Good Luck
Eric

-----Original Message-----
From: Laurent Jacques [mailto:lau...@uc...] 
Sent: Tuesday, July 26, 2011 6:41 AM
To: mat...@li...
Subject: [Matlab-emacs-discuss] fprintf/disp output to another emacs buffer ?

Dear all,

Question: Do you know if it is possible to redirect the output of a matlab disp()/fprintf() (possibly using another command) to another emacs buffer than the one running matlab-shell ?

Objective: When running a "cell" inside a matlab script, this possibility would allow me to display some computation messages or results without to be disturbed by the script code that is always recopied in the matlab shell with the fprintf/disp output.


Best regards,
Laurent




--
Dr. Laurent JACQUES
Postdoctoral Researcher, Return Grant, BELSPO, Belgium

Institute of Information and Communication Technologies,
	Electronics and Applied Mathematics (ICTEAM-TELE) University of Louvain Batiment Stévin Place du Levant 2, PO Box L5.04.04
1348 Louvain-la-Neuve, Belgium

Office: a.133
Tel: +32 10 47 81 24
Fax: +32 10 47 20 89
http://perso.uclouvain.be/laurent.jacques


------------------------------------------------------------------------------
Magic Quadrant for Content-Aware Data Loss Prevention Research study explores the data loss prevention market. Includes in-depth analysis on the changes within the DLP market, and the criteria used to evaluate the strengths and weaknesses of these DLP solutions.
http://www.accelacomm.com/jaw/sfnl/114/51385063/
_______________________________________________
Matlab-emacs-discuss mailing list
Mat...@li...
https://lists.sourceforge.net/lists/listinfo/matlab-emacs-discuss