Re: c-basic-common-init

From: Jeff Mincy <jeff(a)delphioutpost.com&gt; Date: Sat, 10 Nov 2012 17:19:55 -0500 From: Alan Mackenzie <acm(a)muc.de&gt; Date: Sat, 10 Nov 2012 20:45:06 +0000 Hi, Jeff. I got your personal email. Thanks! On Fri, Nov 09, 2012 at 01:24:03PM -0500, jeff-xemacs(a)delphioutpost.com wrote: No. The meaning of (memq 'add-hook-local c-emacs-features) is "The LOCAL parameter to `add-hook' works without first having to call `make-local-hook'.". Ah. The 'or' is correct with that definition. The problem must be the presence of add-hook-local on c-emacs-features as you suggest below. ... This is the case for all modern GNU Emacsen and XEmacs 21.5, but not XEmacs 21.4, I think. `make-local-hook' has been removed from Emacs-24, hence the need to make calling it conditional. The add-hook-local flag is set (in cc-defs.el) by actually testing the effect of (add-hook 'x 'y nil t) on a test buffer. See above. It would seem that CC Mode has got something wrong with either the buffer test or the use of `make-local-hook'. Does 'add-hook-local get added to c-emacs-features in XEmacs 21.4.22? If it does, that is a bug. If it doesn't, why isn't the call to `make-local-hook' working properly? Yes, looks that way. The value of c-emacs-features is: (pps-extended-state add-hook-local gen-string-delim syntax-properties 8-bit) Let me see what I can figure out looking at cc-defs. -jeff For me, on 21.4.22 this code copied from cc-defs returns t: (let ((buf1 (generate-new-buffer " test1")) (buf2 (generate-new-buffer " test2")) changed) (save-excursion ; Needed for XEmacs's byte compiler (set-buffer buf1) (add-hook 'after-change-functions (lambda (beg end old-len) (setq changed t)) nil t) (set-buffer buf2) (insert ?c) (set-buffer buf1) (remove-hook 'after-change-functions (lambda (beg end old-len) (setq changed t)) t) (kill-buffer buf1) (kill-buffer buf2) (not changed))) t This leads to 'add-hook-local being added to c-emacs-features By the way, the add-hook-local value isn't documented in the documentation string for c-emacs-features. However, this code, where I split up the save-excursion returns nil (let ((buf1 (generate-new-buffer " test1")) (buf2 (generate-new-buffer " test2")) changed) (save-excursion ; Needed for XEmacs's byte compiler (set-buffer buf1) (add-hook 'after-change-functions (lambda (beg end old-len) (setq changed t)) nil t)) (save-excursion (set-buffer buf2) (insert ?c) (set-buffer buf1) (remove-hook 'after-change-functions (lambda (beg end old-len) (setq changed t)) t)) (kill-buffer buf1) (kill-buffer buf2) (not changed)) nil My hypothesis is that save-excursion does something that keeps the global after-change-functions from being set until save-excursion exits. Probably gc-pro or something. Maybe somebody else has a better theory, and thus a better fix. I got lost in the unwind protect and insdel C code. It's probably gcpro, since I don't understand that voodoo. -jeff _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://lists.xemacs.org/mailman/listinfo/xemacs-beta