Re: Is xemacs' etags.el really "better and faster"?

From: Vin Shelton <acs(a)xemacs.org&gt; Date: 12 Jun 2001 21:59:39 -0400 Jeff Mincy <jeff(a)delphioutpost.com&gt; writes: > FWIW, I just upgraded to 21.4.3 (more or less painlessly). > 21.4.3 seems zippier and snazzier than 21.1.13, with the exception > of tags lookup. Tag lookup is way slower for me, borderline > annoying. I have tags-build-completion-table turned off. But I > don't think that is causing the problem. 21.1.13 found 'find-tag' > instantly, 21.4.3 took 10-15 seconds to find the same tag and using > the same tag tables. BTW, in 21.1.13, I did find-tag on S\+3 (from ESS), and found the real definition on the third try. 1) (defvar S+elsewhere-dialect-name "S+3" 2) (defvar S+3-dialect-name "S+3" 3) (defun S+3 (&optional proc-name) Whereas 21.4.3 finds the S+3 definition on the first try. So, it's better at finding the right tag, but lots slower.

Snazzier certainly. Zippier I find very hard to believe. There have been numerous complaints about fontifying and about cc-mode. I think the startup code is probably slower in 21.4 also. Can you tell me what you think is faster and tell me what your platform is and what your configuration options are? What feels zippier is just moving around in the code. Perhaps. Well, I have a theory. It seems that lazy lock got turned off in 21.4 and the fontification happened when the file was loaded (and thus the new snazzy progress bar). Because of that, moving around probably was faster because it didn't need to do the lazy font stuff.

I changed my init file to use (turn-on-lazy-lock) instead of the (remove-hook 'font-lock-mode-hook 'turn-on-fast-lock) (add-hook 'font-lock-mode-hook 'turn-on-lazy-lock) Which had been handed down from the custom file from many releases ago. I haven't used cc enough lately to notice any issues. -jeff