> The above already looks wrong: i->left covers 145078 chars, > so i->position should be at least 145078 + 1 (aka BEG). The > difference, 88, is presumably the length of the line you > copied&pasted, so i's position hasn't been updated as it > should have (the `position' field of interval nodes is > updated lazily, so it's sometimes normal for it to be > out-of-date, but not in this case).
Good spot. I had rapidly done C-k C-k C-y C-y to copy a line. I can't remember exactly which line I was copying, but one of the few candidate lines does happen to be 44 chars long. I did C-y C-y, which means I did paste 88 chars (after deleting 44 chars). Of course, the paste was via two inserts, albeit with a very small time interval between them. Maybe I circumvented the lazy update somehow? > > What should I look for re the buffer text properties? The buffer is > > full of them - it's visiting a C++ file. > > Without a reliable way to reproduce the problem, it's going > to be difficult to fix it, because the problem is most likely > introduced sometime *before* we crash. I appreciate that. If someone would like to reproduce my environment, my setup includes: (custom-set-variables ... '(imenu-auto-rescan t) '(imenu-auto-rescan-maxout 1048576) '(imenu-max-items 30) '(imenu-scanning-message nil) '(imenu-sort-function (quote imenu--sort-by-name)) ...) (add-hook 'c-mode-common-hook 'imenu-add-menubar-index) I get the error within a day of editing C++ buffers. However, I'll continue to run under a debugger, and hopefully will be able to provide some useful feedback. Simon. _______________________________________________ emacs-pretest-bug mailing list emacs-pretest-bug@gnu.org http://lists.gnu.org/mailman/listinfo/emacs-pretest-bug