Wrong type argument: number-or-marker-p, nil

[MATSUI Tetsushi]

Since I have switched to python-mode 6, I sometimes encounter the error "Wrong type argument: number-or-marker-p, nil". I could not break it down to minimize the case (and I don't like to open my secret code :) It should be related to indentation, since when it happens I always see the cursor is on column 0 of a new line and no longer TAB works.

Traceback:

Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p nil)
  py-compute-indentation()
  py-newline-and-indent()
  call-interactively(py-newline-and-indent nil nil)

I know this report is too vague, sorry.

[Andreas Roehler (a-roehler)]

Hi,

can't reproduce at the moment.

May you send the output from M-x report-emacs-bug, in case you encounter that next time?

Thanks,

Andreas

[MATSUI Tetsushi (mft)]

The attachment is the main part of the bug report text produced by M-x report-emacs-bug after a wrong type argument error. I hope this helps.

[Alejandro Weinstein (alejandro-weinstein)]

I am observing the same problem. Attached is the output of report-emacs-bug.

[Andreas Roehler (a-roehler)]

need some code example, which triggers the bug.

thanks,

Andreas

[MATSUI Tetsushi (mft)]

Today, when editing the attachment file, I encountered the error on the end of the first class MatrixData.
On the line just after the end of the class definition, hit TAB. This is reproducible on my emacs.

[Andreas Roehler (a-roehler)]

Am 19.12.2011 07:32, schrieb MATSUI Tetsushi:
> Today, when editing the attachment file, I encountered the error on the end of the first class MatrixData.
> On the line just after the end of the class definition, hit TAB. This is reproducible on my emacs.
>
> ** Attachment added: "a file that I encounter the error"
> https://bugs.launchpad.net/python-mode/+bug/901541/+attachment/2639180/+files/container_abc.py
>

hmm, TAB from pos 2348 circles here between 0 and 4, which seems correct.

could you do M-x report-emacs-bug
with this?

[MATSUI Tetsushi (mft)]

I can reproduce the error also with argparse.py of Python 2.7.2, and with python-mode 6.0.3. After opening the file, hit several C-v's until the cursor enters in a class, then hit TAB. Then the error occurs.
The exact position was not the point; in also #7, I used only C-v and <down> to move cursor. If I use C-u C-M-e to move to the end of the class, the error does not occur.

[Andreas Roehler (a-roehler)]

Am 20.12.2011 10:45, schrieb MATSUI Tetsushi:
> I can reproduce the error also with argparse.py of Python 2.7.2, and with python-mode 6.0.3. After opening the file, hit several C-v's until the cursor enters in a class, then hit TAB. Then the error occurs.
> The exact position was not the point; in also #7, I used only C-v and<down> to move cursor. If I use C-u C-M-e to move to the end of the class, the error does not occur.
>

can you send the positions as C-x C-= tells?

[MATSUI Tetsushi (mft)]

The result of C-x = on the line where I reproduce the bug:

Char: SPC (32, #o40, #x20) point=4220 of 87802 (5%) column=0

# C-x C-= is bound to text-style-adjust, which seems irrelevant

Can't you reproduce it still?

[Andreas Roehler (a-roehler)]

file from your attachment here has only

point=3846 of 3845 (EOB) column=0

it's called

container_abc.py

[MATSUI Tetsushi (mft)]

I meant argparse.py in #10 because you are commenting in #9 about my comment #8 which is a report about argparse.py.

[Andreas Roehler (a-roehler)]

Am 20.12.2011 16:38, schrieb MATSUI Tetsushi:
> I meant argparse.py in #10 because you are commenting in #9 about my
> comment #8 which is a report about argparse.py.
>

hmm, Emacs even hangs here with this. Thanks, I'm on it.

[Andreas Roehler (a-roehler)]

does it still occur with last commit -r748 ?

[MATSUI Tetsushi (mft)]

No, it doesn't occur for both container_abc.py and argparse.py.

[MATSUI Tetsushi (mft)]

Though previously reported files do not produce the error (as I said in #15), I am still experiencing this.
I attach the result of report-emacs-bug after the error in ast.py in the Python Library.

[MATSUI Tetsushi (mft)]

I had to add more info.

The version is revision 756 of bzr repository.

If you need, the result of C-x = is:
Char: SPC (32, #o40, #x20) point=10561 of 11805 (89%) column=0

[Georg Brandl (gbrandl)]

I'm also seeing this when pressing RET in a docstring. This is a debug trace:

Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p nil)
  <(nil 6633)
  (if (< (nth 2 (if (featurep (quote xemacs)) (parse-partial-sexp (point-min) (point)) (syntax-ppss))) (line-beginning-position)) (current-indentation) (ignore-errors (goto-char (nth 2 pps))) (py-line-backward-maybe) (back-to-indentation) (py-compute-indentation orig origline closing line inside repeat))
  (progn (forward-line -1) (end-of-line) (skip-chars-backward "
\n\f") (if (< (nth 2 (if (featurep (quote xemacs)) (parse-partial-sexp (point-min) (point)) (syntax-ppss))) (line-beginning-position)) (current-indentation) (ignore-errors (goto-char (nth 2 pps))) (py-line-backward-maybe) (back-to-indentation) (py-compute-indentation orig origline closing line inside repeat)))
  (if (eq origline (py-count-lines)) (progn (forward-line -1) (end-of-line) (skip-chars-backward "
\n\f") (if (< (nth 2 (if (featurep (quote xemacs)) (parse-partial-sexp (point-min) (point)) (syntax-ppss))) (line-beginning-position)) (current-indentation) (ignore-errors (goto-char (nth 2 pps))) (py-line-backward-maybe) (back-to-indentation) (py-compute-indentation orig origline closing line inside repeat))) (ignore-errors (goto-char (nth 2 pps))) (py-line-backward-maybe) (back-to-indentation) (py-compute-indentation orig origline closing line inside repeat))
  (cond ((and (bobp) (eq origline (py-count-lines))) (current-indentation)) ((and (bobp) (py-statement-opens-block-p)) (+ (if py-smart-indentation (py-guess-indent-offset nil orig) py-indent-offset) (current-indentation))) ((and (bobp) (not (py-statement-opens-block-p))) (current-indentation)) ((and (nth 3 pps) (nth 8 pps)) (if (eq origline (py-count-lines)) (progn (forward-line -1) (end-of-line) (skip-chars-backward "
\n\f") (if (< (nth 2 (if ... ... ...)) (line-beginning-position)) (current-indentation) (ignore-errors (goto-char (nth 2 pps))) (py-line-backward-maybe) (back-to-indentation) (py-compute-indentation orig origline closing line inside repeat))) (ignore-errors (goto-char (nth 2 pps))) (py-line-backward-maybe) (back-to-indentation) (py-compute-indentation orig origline closing line inside repeat))) ((and (looking-at "\"\"\"\\|'''") (not (bobp))) (py-beginning-of-statement) (py-compute-indentation orig origline closing line inside repeat)) ((nth 8 pps) (if (eq origline (py-count-lines)) (progn (goto-char (nth 8 pps)) (py-line-backward-maybe) (skip-chars-backward " ") (py-compute-indentation orig origline closing line inside repeat)) (goto-char (nth 8 pps)) (if (and line (or py-indent-honors-inline-comment (looking-back "^[ ]*"))) (current-column) (forward-char -1) (py-compute-indentation orig origline closing line inside repeat)))) ((and (looking-at "[ ]*#") (looking-back "^[ ]*") (not (eq (line-beginning-position) (point-min)))) (forward-line -1) (end-of-line) (setq line t) (py-compute-indentation orig origline closing line inside repeat)) ((nth 1 pps) (cond ((and inside (not line)) (when (and (eq (point) orig) (looking-at "[ ]*\\()\\)[ ]*$")) (setq closing (match-beginning 0))) (save-excursion (goto-char (nth 1 pps)) (setq this-li...

[Andreas Roehler (a-roehler)]

Am 02.01.2012 14:50, schrieb Georg Brandl:
> I'm also seeing this when pressing RET in a docstring. This is a debug
> trace:

get the error it from pos 316, inside tqs, thanks

[Andreas Roehler (a-roehler)]

please re-open, if the bug still exists

[Georg Brandl (gbrandl)]

Fixed it for my case, thanks!

[MATSUI Tetsushi (mft)]

Error doesn't occur for me now, thanks.