Hi, I remember having trouble getting the below to work when the text being matched was split between multiple calls to the process filter. Perhaps I erroneously ended up with something that fit Emacs' timing, but not XEmacs? Also, I think you have an older version of matlab.el. Try out the latest version from CVS and perhaps that will help. Thanks Eric -----Original Message----- From: Uwe Brauer [mailto:oub.oub.oub＠gmail.com] On Behalf Of Uwe Brauer Sent: Monday, January 10, 2011 8:42 AM To: XEmacs Beta Discussion Cc: Eric Ludlam Subject: problems with make-extent in matlab.el using the matlab debugger. Hello I am not sure whether the problem I describe has to do with the implementation of make-extent in xemacs or the problems lies in a different part of the code. Matlab.el (version 3.1.1) allows to debug matlab code. During the debugging matlab.el, split the window, opens the file you debug, the cursor jumps from line to line and show in the other window via the matlab-shell the result of the executed command. However GNU emacs and Xemacs behave differently In GNU emacs you see in the matlab-shell buffer Someting like 1 tspan=[0 1] Where 1 is underlined and clicking with the mouse on it is jumps to the matlab file which is debugged. In Xemacs however the matlab-buffer displays <a href="matlab: opentoline('/home/oub/ALLES/Matlab-init/testeuler.m',1,1)">1 </a>tspan=[0 1] Which I find very annoying. I tried to debug it but can localise the problem. It seems to me the issue is the code (defun matlab-shell-render-errors-as-anchor (str) "Detect non-url errors, and treat them as if they were url anchors. Argument STR is the text that might have errors in it." (when (string-match matlab-shell-error-stack-end str) (save-excursion (when (re-search-backward matlab-shell-error-stack-start nil t) (let ((groupend nil) (first nil)) (save-excursion (setq groupend (re-search-forward matlab-shell-error-stack-end)) ) ;; We have found an error stack to investigate. (while (re-search-forward gud-matlab-error-regexp nil t) (let* ((err-start (match-beginning 0)) (err-end (match-end 0)) (err-text (match-string 0)) (err-file (match-string 2)) (err-line (match-string 3)) (o (matlab-make-overlay err-start err-end)) (url (concat "opentoline('" err-file "'," err-line ",0)")) ) (matlab-overlay-put o 'mouse-face 'highlight) (matlab-overlay-put o 'face 'underline) ;; The url will recycle opentoline code. (matlab-overlay-put o 'matlab-url url) (matlab-overlay-put o 'keymap matlab-shell-html-map) ;; Keep track of the very first error in this error stack. ;; It will represent the "place to go" for "go-to-last-error". (matlab-overlay-put o 'first-in-error-stack first) (when (not first) (setq first url)) )) ))))) Where the various overlay functions are defaliases to (defalias 'matlab-make-overlay 'make-extent) (defalias 'matlab-overlay-put 'set-extent-property) (defalias 'matlab-overlay-get 'extent-property) (defalias 'matlab-delete-overlay 'delete-extent) (defalias 'matlab-overlay-start 'extent-start-position) (defalias 'matlab-overlay-end 'extent-end-position) (defalias 'matlab-previous-overlay-change 'previous-extent-change) (defalias 'matlab-next-overlay-change 'next-extent-change) Etc. So since GNU emacs and Xemacs behave so differently I thought that maybe make-extent and friend a implemented in a different way. Anybody has an idea? regards Uwe Brauer _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://lists.xemacs.org/mailman/listinfo/xemacs-beta