max-specpdl-size error in mumamo-do-fontify

[ecce berlin]

I really like nxhtml-mode, but it's not working in GNU Emacs 23.0.60.1 (the current "emacs-snapshot-gtk" in Ubuntu-8.10). When I open ANY file that wants to use nxhtml-mode (.html, .php etc.), I get (something like) the following error in the mini-buffer:

mumamo-do-fontify-2: mumamo-do-fontify m=nxhtml-mode, s/e=1618/2201 syn-min/max=1615/2589: Variable binding depth exceeds max-specpdl-size

Emacs gets irresponsive, sometimes even crashes.

I've attached the *Messages* buffer with the errors encountered.

I've also attached the output of nxhtmltest-run (emacs was invoked with -q) and the Messages-buffer it filled (quite large!).

Here the versions I use:
GNU Emacs 23.0.60.1
nxhtml-1.60-081006

[lborgman (lennart-borgman)]

Hi ecce,

Thanks for the error report. Sorry for taking some time before looking at this.

It looks like there is some mismatch between nXhtml and either your emacs init files or the developing version of Emacs you are using. Could you please try the command

  M-x nxhtmltest-run-Q

so we can find out where to start looking for the error?

[ecce berlin (ecce-berlin)]

Hello,

Thanks for your reply!

M-x nxhtmltest-run-Q seems to successfully run some tests, but all it gives back is the following message in the minibuffer:
Wrong type argument: stringp, nil

*Messages* says:
Test message
Ran 30 tests, 30 results were as expected
Wrote /home/ecce/tmp/lisp/nxhtml/tests/temp-test.el
nxhtmltest-bin-Q=/home/ecce/tmp/lisp/nxhtml/tests/
nxhtml-auto-start=/home/ecce/tmp/lisp/nxhtml/autostart.el
After set nxhtmltest-run-Q=run
let*: Wrong type argument: stringp, nil

And the debugger:
Debugger entered--Lisp error: (wrong-type-argument stringp nil)
  call-process(nil nil 0 nil "-Q" "-l" "/home/ecce/tmp/lisp/nxhtml/tests/temp-test.el" "-l" "/home/ecce/tmp/lisp/nxhtml/autostart.el" "-l" "/home/ecce/tmp/lisp/nxhtml/tests/nxhtmltest-suites.el")
  (let* ((test-el ...) (nxhtml-auto-start ...) (temp-eval-file ...) (temp-eval-buf ...) (load-path load-path)) (add-to-list (quote load-path) nxhtmltest-bin-Q) (require (quote nxhtmltest-helpers)) (nxhtmltest-get-fontification-method) (with-current-buffer temp-eval-buf (erase-buffer) (insert "(setq debug-on-error t)\n" "(eval-when-compile (require 'cl))\n" "(delete-other-windows)\n" "(eval-after-load 'nxhtml '(setq nxhtml-skip-welcome t))\n" ...)) (when (featurep ...) (with-current-buffer temp-eval-buf ...)) (with-current-buffer temp-eval-buf (save-buffer)) (unless (file-exists-p nxhtmltest-bin-Q) (error "Can't find directory %s" nxhtmltest-bin-Q)) (unless (file-exists-p test-el) (error "Can't find file %s" test-el)) (unless (file-exists-p nxhtml-auto-start) (error "Can't find file %s" nxhtml-auto-start)) (message "nxhtmltest-bin-Q=%s" nxhtmltest-bin-Q) (message "nxhtml-auto-start=%s" nxhtml-auto-start) (setenv "nxhtmltest-run-Q" "run") (message "After set nxhtmltest-run-Q=%s" (getenv "nxhtmltest-run-Q")) (call-process (ourcomments-find-emacs) nil 0 nil "-Q" "-l" temp-eval-file "-l" nxhtml-auto-start "-l" test-el) (message "After call-process") (setenv "nxhtmltest-run-Q") (message "Starting new Emacs instance for test - it will be ready soon ..."))
  nxhtmltest-run-Q()
  call-interactively(nxhtmltest-run-Q t nil)
  execute-extended-command(nil)
  call-interactively(execute-extended-command nil nil)

At that time, there's a buffer "temp-test.el" with the following contents:
(setq debug-on-error t)
(eval-when-compile (require 'cl))
(delete-other-windows)
(eval-after-load 'nxhtml '(setq nxhtml-skip-welcome t))
(setq nxhtmltest-default-fontification-method 'fontify-w-timer-handlers)

One last thing I noticed: If I run the command a second time, I additionally get this error:
ad-Orig-syntax-ppss-flush-cache: Lisp nesting exceeds `max-lisp-eval-depth'

Hope this helps...

[ecce berlin (ecce-berlin)]

Hm... actually, I get a very similar result for nxhtmltest-run-Q with this version of Emacs, which otherwise seems to work fine with nxhtml:

GNU Emacs 22.1.1 (i486-pc-linux-gnu, GTK+ Version 2.12.9)
of 2008-05-03 on terranova, modified by Ubuntu

*Messages* says:
Loading /home/ecce/tmp/lisp/nxhtml/tests/nxhtmltest-Q.el (source)...done
Loading cl-seq...done
Loading cl-extra...done
Test message
Loading pp...done
Ran 30 tests, 30 results were as expected
Wrote /home/ecce/tmp/lisp/nxhtml/tests/temp-test.el
nxhtmltest-bin-Q=/home/ecce/tmp/lisp/nxhtml/tests/
nxhtml-auto-start=/home/ecce/tmp/lisp/nxhtml/autostart.el
After set nxhtmltest-run-Q=run
let*: Wrong type argument: stringp, nil

And the debugger:
Debugger entered--Lisp error: (wrong-type-argument stringp nil)
  call-process(nil nil 0 nil "-Q" "-l" "/home/ecce/tmp/lisp/nxhtml/tests/temp-test.el" "-l" "/home/ecce/tmp/lisp/nxhtml/autostart.el" "-l" "/home/ecce/tmp/lisp/nxhtml/tests/nxhtmltest-suites.el")
  (let* ((test-el ...) (nxhtml-auto-start ...) (temp-eval-file ...) (temp-eval-buf ...) (load-path load-path)) (add-to-list (quote load-path) nxhtmltest-bin-Q) (require (quote nxhtmltest-helpers)) (nxhtmltest-get-fontification-method) (with-current-buffer temp-eval-buf (erase-buffer) (insert "(setq debug-on-error t)\n" "(eval-when-compile (require 'cl))\n" "(delete-other-windows)\n" "(eval-after-load 'nxhtml '(setq nxhtml-skip-welcome t))\n" ...)) (when (featurep ...) (with-current-buffer temp-eval-buf ...)) (with-current-buffer temp-eval-buf (save-buffer)) (unless (file-exists-p nxhtmltest-bin-Q) (error "Can't find directory %s" nxhtmltest-bin-Q)) (unless (file-exists-p test-el) (error "Can't find file %s" test-el)) (unless (file-exists-p nxhtml-auto-start) (error "Can't find file %s" nxhtml-auto-start)) (message "nxhtmltest-bin-Q=%s" nxhtmltest-bin-Q) (message "nxhtml-auto-start=%s" nxhtml-auto-start) (setenv "nxhtmltest-run-Q" "run") (message "After set nxhtmltest-run-Q=%s" (getenv "nxhtmltest-run-Q")) (call-process (ourcomments-find-emacs) nil 0 nil "-Q" "-l" temp-eval-file "-l" nxhtml-auto-start "-l" test-el) (message "After call-process") (setenv "nxhtmltest-run-Q") (message "Starting new Emacs instance for test - it will be ready soon ..."))
  nxhtmltest-run-Q()
  call-interactively(nxhtmltest-run-Q)
  execute-extended-command(nil)
  call-interactively(execute-extended-command)

[lborgman (lennart-borgman)]

I think this is a bug I have fixed in the beta. Can you please try the latest beta here:

  http://ourcomments.org/Emacs/DL/elisp/nxhtml/beta/

But do not through away the version of nXhtml you are using, the beta is not ready for use yet.

[ecce berlin (ecce-berlin)]

Hello,

Thank you! I tried the beta and there is no "mumamo-do-fontify" error in the mini-buffer anymore. However, I get lots of these errors instead, when moving through the document:

start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [3 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [32 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [3 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [7 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [3 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [7 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [3 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [7 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [110 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [113 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil
ERROR: new-way /= old-way
start-in-cons=nil [43 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [10 times]
pos=1962, old-chunk=#<overlay from 1962 to 1985 in test.php>
start-in-cons=nil [3 times]
pos=1985, old-chunk=#<overlay from 1985 to 2035 in test.php>
start-in-cons=nil [3 times]
pos=2035, old-chunk=#<ov...

[lborgman (lennart-borgman)]

Thanks ecce,

Sorry for the extra messages, forgot to remove them. I will upload a new version when I have fixed some problems. You can of course comment out those messages yourself if you know how to do it. (Comments in elisp starts with a semi-colon. The code that gives the messages are in mumamo.el.)

It looks to me like you are still running the old version of nxhtmltest-run-Q. Can you do

  C-h f nxhtmltest-run-Q

There is a link in the help buffer to the code. Click on that link. Does it go to the new nXhtml beta code?

[ecce berlin (ecce-berlin)]

Thank you!

The link in the help buffer of nxhtmltest-run-Q does go to the beta code (in the nxhtml beta directory: nxhtml/tests/nxhtmltest-Q).

I noticed (using the beta) that the error messages differ depending on the type of file I open:

For a pure html file I still get this:
mumamo-do-fontify-2: mumamo-do-fontify m=nxhtml-mode, s/e=602/799 syn-min/max=1/1126: Variable binding depth exceeds max-specpdl-size
MuMaMo error, please look in the *Message* buffer
start-in-cons=nil [12 times]
mumamo-do-fontify-2: mumamo-do-fontify m=nxhtml-mode, s/e=547/602 syn-min/max=1/1126: Variable binding depth exceeds max-specpdl-size

For a php file (with css and javascript parts) I get this:
start-in-cons=(580 javascript-mode (nxml-mode))
start-in-cons=nil [36 times]
ad-Orig-syntax-ppss-flush-cache: Variable binding depth exceeds max-specpdl-size
start-in-cons=nil
start-in-cons=(1415 css-mode (nxml-mode))
start-in-cons=nil [2 times]
Internal error in rng-validate-mode triggered at buffer position 1367. Wrong type argument: number-or-marker-p, nil

Both still have "Variable binding depth exceeds max-specpdl-size", but in the latter case with "ad-Orig-syntax-ppss-flush-cache".

[lborgman (lennart-borgman)]

Thanks for the additional information.

I would like you to try the command

   M-x emacs-Q-nxhtml

but I do not think it will work for you at the moment since nxhtmltest-run-Q does not work. Let us try to find out what is wrong there first.

Can you tell me what

  C-h v exec-directory

gives? Is your emacs executable there? What is the name of it?

[ecce berlin (ecce-berlin)]

Hello and thanks for your time!

"emacs-Q-nxhtml" didn't work, same error again.
let: Wrong type argument: stringp, nil

Then I checked "exec-directory": Its value is
"/usr/lib/emacs/23.0.60/i486-linux-gnu/"

However, there is in fact no executable emacs binary there!

To circumvent the problem, I created there a soft link named "emacs" to
the actual binary: /usr/bin/emacs

Now nxhtmltest-run-Q works. Its output is attached.

[lborgman (lennart-borgman)]

Thanks, that is very good. I have an old unconfirmed bug report regarding max-spepdl-size that is unconfirmed. Maybe we can find out what is wrong here. What is the value of this variable on your system?

Can you in the new Emacs instance started with nxhtmltest-run-Q increase the value of this variable and then do

   M-x nxhtmltest-run

(without the "-Q" at the end).

Though I wonder why the genshi tests fails too. I have no idea right now.

[ecce berlin (ecce-berlin)]

max-specpdl-size is set to 1000.
I did as you said and increased it to 3000, but trying to run nxhtmltest-run immediately gave the error:

Lisp error: (error "Lisp nesting exceeds `max-lisp-eval-depth'")

max-lisp-eval-depth is set to 400. So I also increased it to 3000 and tried again.
This time nxhtmltest-run finished with the output attached!

[lborgman (lennart-borgman)]

Thanks again. I do not understand why this happens to you but not to me. What does

   M-x version

say, it should give the build date (which is perhaps the same as the checkout date ...).

To begin with I am a bit curious about indent-bug290364, Can you put point on nxhtml-ert-indent-bug290364 and press "d"? This should give a backtrace I would like to see.

[ecce berlin (ecce-berlin)]

Hi again,

I might have good news: You suggested an incompatibility with the emacs version I'm using, so I downloaded the current CVS version today and compiled it myself:

GNU Emacs 23.0.60.1 (i686-pc-linux-gnu, GTK+ Version 2.12.9)
 of 2008-11-29 on evo

instead of the Ubuntu package:
https://launchpad.net/ubuntu/intrepid/+source/emacs-snapshot/1:20081013-1
GNU Emacs 23.0.60.1 (i486-pc-linux-gnu, GTK+ Version 2.14.3)
 of 2008-10-13 on rothera, modified by Debian

And in fact, nxhtmltest-run-Q does much better (output is attached)!

So maybe it wasn't a nxhtml issue at all but a temporary bug in emacs?

Still, I'm encountering the following two errors, but they seem much less severe:

Internal error in rng-validate-mode triggered at buffer position 1853. Wrong type argument: number-or-marker-p, nil

ERROR: new-way /= old-way

[ecce berlin (ecce-berlin)]

Hello,

I attached the backtrace for nxhtml-ert-indent-bug290364.

[Chris Conway (cconway)]

I'm having a similar error with version 1.61-081128-08_06_51 downloaded from http://ourcomments.org/Emacs/DL/elisp/nxhtml/beta/

I get the following message when I open the attached HTML file:
mumamo-do-fontify-2: mumamo-do-fontify m=nxhtml-mode, s/e=542/822 syn-min/max=1/822: Variable binding depth exceeds max-specpdl-size

In addition to the error, the entire document is highlighted in pale yellow and there is no syntax highlighting.

The error goes away and syntax highlighting works correctly if I remove the DOCTYPE declaration from the beginning of the file.

Here's my Emacs version info:
GNU Emacs 23.0.60.1 (i486-pc-linux-gnu, GTK+ Version 2.14.3) of 2008-10-13 on rothera, modified by Debian

[Chris Conway (cconway)]

Also note: I had the same error with nxhtml versions 1.59 and 1.60 (same version of Emacs).

[lborgman (lennart-borgman)]

Chris and ecce,

I need your help. Can you please test the new beta I just uploaded?

[Chris Conway (cconway)]

I'm still seeing the same behavior with nxhtml-1.61.20081219-03.

With the above-attached HTML file, I get the same error, same pale yellow background, and the modeline:
(nXhtml/nxhtml Invalid L hs Outl)

Executing M-x nxhtml-mode fixes the syntax coloring, and the modeline becomes:
(nXhtml Invalid L hs Outl)

Deleting the DOCTYPE declaration actually causes the file to open up in html-helper-mode instead of nxhtml-mode.

[lborgman (lennart-borgman)]

Hi Chris, thanks for the feedback! I can however not see this problem.

Did you try

  M-x emacs-Q-nxhtml

to start a new Emacs instance and then opening the file from there?

[Chris Conway (cconway)]

M-x emacs-Q-nxhtml fails with
    let: Wrong type argument: stringp, nil

[lborgman (lennart-borgman)]

Thanks Chris, then it looks like there is a bug in the development version of Emacs you are using that is not there in the version I am using.

Is there any chance you can try with a newer version? (You said above you are using a version from 2008-10-13.)

[Chris Conway (cconway)]

The above-reported results are for nxhtml-1.61.20081219-03.

I get the same highlighting behavior and the same emacs-Q-nxhtml failure with Emacs 22.2.1, though not the mumamo-do-fontify-2 error.

[lborgman (lennart-borgman)]

Thanks.
I think I misunderstood you. I missed that you can still not start with

   M-x emacs-Q-nxhtml

Can you please download the latest version I just uploaded and then turn on debug-on-error ("Enter Debugger on Error" in the Options menu) and try the command above again?

[Chris Conway (cconway)]

With nxhtml-1.61-081220-02_06_26, M-x emacs-Q-nxhtml works. Oddly, I get the mumamo-do-fontify error in the session opened up by M-X emacs-Q-nxhtml, but not in the original session! The fontification problem is the same in both instances, and seems to be a separate issue.

I'm attaching the output in *Messages* from the emacs-Q-nxhtml instance.

[lborgman (lennart-borgman)]

What is the value of max-specpdl-size?

   C-h v max-specpdl-size

Can you try to double it?

   M-: (setq max-specpdl-size (* 2 max-specpdlsize))

[Chris Conway (cconway)]

The value of max-specpdl-size is 1000. I set both max-specpdl-size and max-lisp-eval-depth to 100000 and still got the same error.

[lborgman (lennart-borgman)]

And that is from "M-x emacs-Q-nxhtml" I suppose?

[lborgman (lennart-borgman)]

Hi Chris, I wonder if this is the same problem that was cured for ecce berlin when he upgraded from CVS Emacs 2008-10-13 to a later version? (See the message above from ecce on 2008-11-29.)

[Chris Conway (cconway)]

Right, I did

    M-x emacs-Q-nxhtml

then, in the new instance

   M-x set-variable max-specpdl-size 100000
   M-x set-variable max-lisp-eval-depth 100000
   C-c C-f index.html

(where index.html is the file attached above) and I still get the max-specpdl-size error.

I am using the package emacs-snapshot=1:20081013-1 from ubuntu, so it's quite possible my problem is the same as ecce's and will go away when I upgrade. I'm installing a PPA from https://launchpad.net/~ubuntu-elisp/+archive right now and I'll update you if it fixes the problem.

[lborgman (lennart-borgman)]

Fine, please update me if it does not work too ;-)

[Chris Conway (cconway)]

I'm seeing the same behavior with emacs-snapshot=1:20081205-1 installed from the PPA. I.e., after M-x emacs-Q-nxhtml, I still get a max-specpdl-size or max-lisp-eval-depth error, no matter how high I raise the values. M-x emacs-version returns:

    GNU Emacs 23.0.60.1 (i486-pc-linux-gnu, GTK+ Version 2.14.4) of 2008-12-05 on iridium, modified by Debian

[lborgman (lennart-borgman)]

I am not able to understand what is going on. I will try to get some help from Emacs Devel list.

Ecce, could you please try the file Chris has problem with and see if it works for you?

[lborgman (lennart-borgman)]

Chris, I got an answer on Emacs developers list saying that it could depend on which files have been byte-compiled:

    http://lists.gnu.org/archive/html/emacs-devel/2008-12/msg00792.html

Is it possible that you can check out Emacs yourself from the CVS repository, compile it and install it?

[Chris Conway (cconway)]

I built Emacs from source and the max-specpdl-error went away. On startup, emacs gives the warning

    Warning (initialization): Building Emacs overflowed pure space. (See the node Pure Storage in the Lisp manual for details.)

I'm attaching the output of M-x nxhtmltest-run

[Fabián Ezequiel Gallina (fgallina)]

I have exactly the same problem and I am running the weekly snapshot for Debian, the version is: GNU Emacs 23.0.60.1 (x86_64-pc-linux-gnu, GTK+ Version 2.12.11) of 2008-12-21 on elegiac, modified by Debian

I Tested also with the build of 2008-12-05 both on Debian Lenny and Ubuntu 8.10

I tried with the Beta provided above, the latest release and the trunk. With all of them I still receiving this error.

Attached is my *Messages* buffer, I hope this helps but I'm fairly conviced this has to be with an Emacs regression since it works fine the version 22 and as if I recall correctly it worked on previous version of Emacs 23.

[Fabián Ezequiel Gallina (fgallina)]

Good news, I compiled the latest cvs version of Emacs and fontification worked fine with both nxhtml-1.60-081006 and nxhtml-1.61-081222-00_59_41 so I guess users which use Emacs from the weekly snapshots will have to wait to the next build after 2008-12-21.

[lborgman (lennart-borgman)]

Thanks Fabian. I will ask Romain if he can build a new snap shot.

Did you see any problem with pure space overflow when you built Emacs?

[Fabián Ezequiel Gallina (fgallina)]

Thanks to you for such amazing package :).

Fortunately I don't have that pure space overflow problem, I suggest disabling all packages but nxhtml and see what happens, perhaps that bug is from another package.

By the way I tested that pure space overflow problem with the 1.61 beta version and 1.60 release.

[lborgman (lennart-borgman)]

Thanks Fabian, I am glad you like it.

The pure space overflow message is something you see at start up of Emacs in the *Messages* buffer. Could you start with

  emacs -Q

and then look in the message buffer? You do that with

  C-h e

Do you see a message similar to that Chris found above 2008-12-22?

[Fabián Ezequiel Gallina (fgallina)]

I thought the error was in the startup by letting Emacs read the .emacs file, it doesn't matters if I let Emacs to load the .emacs file or not (-Q), I have not errors or warnings listed on my *Messages* buffer.

[Chris Conway (cconway)]

To be clear, he "space overflow" message shows up in *Warnings*, not *Messages* (which is empty), and shows up with both "emacs -Q" and just plain "emacs". (I get unrelated errors when I load my .emacs, because the fresh build doesn't find some system elisp packages.)

[Fabián Ezequiel Gallina (fgallina)]

Chris, none *Warnings* buffer is created in my Emacs at startup (with and without -Q option). Did you tried to get again the latest cvs version and compile it again? I think that should do the trick. Otherwise you could wait till the next release of the weekly snapshots, I think it will be solved for sure.

[Chris Conway (cconway)]

Fabián, I'm sure it's an unrelated regression. It's occuring at the
latest CVS version, but not in the last snapshot.

On Wed, Dec 24, 2008 at 12:12 AM, Fabián Ezequiel Gallina
<email address hidden> wrote:
> Chris, none *Warnings* buffer is created in my Emacs at startup (with
> and without -Q option). Did you tried to get again the latest cvs
> version and compile it again? I think that should do the trick.
> Otherwise you could wait till the next release of the weekly snapshots,
> I think it will be solved for sure.
>
> --
> max-specpdl-size error in mumamo-do-fontify
> https://bugs.launchpad.net/bugs/300946
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[lborgman (lennart-borgman)]

Chris, could you please file an Emacs bug report for the pure space overflow? You do that with

  M-x report-emacs-bug

It might otherwise get overlooked and it is best if you do it since the person fixing it probably want you to test again.

[Chris Conway (cconway)]

Lennart, I just sent the bug report.

[lborgman (lennart-borgman)]

Thanks Chris.

[Fabián Ezequiel Gallina (fgallina)]

Hi, I have downloaded http://ourcomments.org/Emacs/DL/elisp/nxhtml/zip/nxhtml-1.75-090112.zip and installed on my emacs, I'm using Emacs from the weekly snapshots and I'm getting the mumamo-do-fontify-2: mumamo-do-fontify m=nxhtml-mode, s/e=1509/1544 syn-min/max=1509/1557: Variable binding depth exceeds max-specpdl-size error.

It's very funny because it didn't happened with a own compiled version but it happens with the precompiled version of emacs from the weekly snapshots. Also I tested with wine the version provided on the nxhtml website and it worked without that error. I think we should find the cause for this issue.

Attached is my *Messages* buffer.

The version of Emacs is:

GNU Emacs 23.0.90.1 (x86_64-pc-linux-gnu, GTK+ Version 2.12.12) of 2009-02-22 on elegiac, modified by Debian

Also for the comment the behavior with my compiled version of emacs was way too buggy too.

[hobbes (hobbes-poukram)]

Hello all,

I'm also stopped in my usage of nxhtml by this error on debian. Has any progress been made ? Is there anything we can do to help ?

Thanks

[Christoph-D (eigene-homepage)]

Hello,

why does this bug report have the status "fix released"? There seem to be many people around still affected by it.

I tried to install nxhtml on Ubuntu 8.10/x86 with the emacs-snapshot package from the ubuntu repository and with the more recent emacs-snapshot package from https://launchpad.net/~ubuntu-elisp/+archive/ppa .

In both cases I get the error "Variable binding depth exceeds max-specpdl-size". First it only shows in the minibuffer but when I enter a character in a <?php ?> block, the lisp debugger pops up with a backtrace (see attachment). It makes no difference if I start emacs with -Q or not.

The attached backtrace is from the weekly emacs-snapshot package. M-x version says:
GNU Emacs 23.0.91.1 (i486-pc-linux-gnu, GTK+ Version 2.14.4) of 2009-03-21 on hassium, modified by Debian

Best regards,
Christoph

[primewizard (prime-wizard+launchpad)]

I have the same problem
but it't ok in MS Win XP

my temporary solution:

(setq ad-redefinition-action 'discard) ;; <-- insert
(load "~/.emacs.d/nxhtml/autostart.el")

my environment:
Debian Lenny
GNU Emacs 23.0.91.1
nXhtml mode version 1.76

[Fabián Ezequiel Gallina (fgallina)]

Guys, I'm quite sure this is not really a nxhtml bug since it works well when I compile Emacs myself.

Perphaps this is a problem with the weekly snapshots.

[Mathias Schenner (mathias.schenner)]

Same here and for ecce above: https://bugs.launchpad.net/nxhtml/+bug/300946/comments/16

Something in the Debian packaging process seems to cause the conflict,
even in the recent jaunty snapshots.

[Angus77 (andrewmcintosh-yahoo)]

I can confirm this. No matter what .deb version of emacs-snapshot I use, I get this problem. But when I compiled it myself this morning (23.0.92.1), the problem disappeared.

[Dennis de Vaal (d-devaal)]

I also can confirm this. Unfortunately the problem still persists (Variable binding depth exceeds max-specpdl-size) while opening *.php files, with the latest emacs-snapshot from emacs.orebokech.com version: 23.0.92.1 with nxhtml- 1.76-090226. I really would like to use nxhtml but this makes it impossible, and it seems the problem is still not fixed. I haven't tried compiling emacs from cvs, but that's not real solution anyway.

[lborgman (lennart-borgman)]

I have contacted Romain Francoise (who seems to make the snap-shots for Ubuntu) and asked if he can find out what is wrong.

[lborgman (lennart-borgman)]

Could people who have this problem please try the latest beta that I just uploaded. (See EmacsWiki.)

I have got some contact with Romain and this is our conversation so far:

>>> I see. Which packages are loaded as add-ons?
>>
>> All the distribution packages based on emacsen-common.
>>
>>> Where are they loaded, in site-start.el, or?
>>
>> Debianized Emacsen load /usr/share/emacs/site-lisp/debian-startup.el
>> which in turns loads all the files in /etc/emacs/site-start.d and
>> /etc/<emacs-flavor>/site-start.d.
>
> Thanks. I do not have Ubuntu here (sigh), but I downloaded the files.
>
> Loading debian-startup.el does not run any code so I guess you mean
> that the function debian-startup in that file is run after loading the
> file, or?
>
> I can't see that running debian-startup does anything at all with a
> default installation. Am I missing something?
>
> I guess debian has patched startup.el? Anything more? How does the patch look?

[Dennis de Vaal (d-devaal)]

I tried nxhtml-1.77-090430-02_30_54 and the problem doesn't appear anymore. Thanks Iborgman!

[lborgman (lennart-borgman)]

Thanks Dennis, this is good. I can probably release 1.77 very soon then.

[Angus77 (andrewmcintosh-yahoo)]

Fixed it here, too. Many thanks! This bug was maddening!

[lborgman (lennart-borgman)]

Thanks Angus77. Sorry that I could not fix it earlier.

[neomantic (calbers)]

I'm not sure this bug was really fixed. I get the same "mumamo-do-fontify-2" errors as the original poster using nxhtml-1.77-090509-18_54_38.zip and from http://emacs.orebokech.com/ on my debian sid. The version of this snapshot is
GNU Emacs 23.0.93.1 (i486-pc-linux-gnu, GTK+ Version 2.16.1)
 of 2009-05-01 on elegiac, modified by Debian

Any assistance would be great.

[lborgman (lennart-borgman)]

Hi neomatic,

Some people have found that the snap-shots currently does not work with nXhtml, but a build directly from Emacs cvs repository does. I am trying to find out what the problem is.

Could you perhaps try and compile Emacs yourself and see if you still have a problem? (BTW I have uploaded a new beta 1.77 today.)

[neomantic (calbers)]

Hi,

Yes, indeed, the mumamo-do-fontify errors do go away whenc emacs 23.0.93.1 is compiled from the cvs sources. So there's something up with the debian snapshot package.

[lborgman (lennart-borgman)]

Thanks neomantic. However there seem to be a solution using the snapshots now. Can you please look at https://bugs.launchpad.net/nxhtml/+bug/369800 and try the same thing that I have suggested some hours ago to hobbes there?

[neomantic (calbers)]

Indeed, that works. Thanks. I am, now, of course getting this bug: https://bugs.launchpad.net/nxhtml/+bug/354363. Tidy cleanup is also failing for me...another bug. Sorry, the software is too buggy for me. I'll check in on the project later. Good luck.

[lborgman (lennart-borgman)]

Thanks neomantic.

It would be very helpful if you told more about the bug you are now seeing. Are you saying that if you are opening the file test.php that hobbes supplied you get the error in https://bugs.launchpad.net/nxhtml/+bug/354363 or do you get a similar error (but that is perhaps a very different error)?

And can you please report the error in tidy? I have no idea what error you get and under what circumstances.

[lborgman (lennart-borgman)]

I have not get any answer back from neomantic, but I wonder if the problem with Tidy was the same as seen in this question:

  https://answers.launchpad.net/nxhtml/+question/85493

In that case it is fixed in nxhtml ver 1.96