Re: [CEDET-devel] semantic-analyze-current-context bug
Brought to you by:
zappo
From: peng yu <yup...@gm...> - 2009-04-19 08:33:54
|
Hi, Eric. I test your patch on some of my code, it seems work very well. The command semantic-analyze-current-context will never left some file opened. And I also test some other functions such as smart completion, all of them work correct. Best regard. YuPeng 2009/4/18 Eric M. Ludlam <er...@si...> > > Hi, >> >> Ah, it was in the symref code. Below is a small patch. The global >> database was explicitly asking for symref to keep the buffer open. I >> don't remember why, and I didn't leave comments behind to explain to >> myself why. >> >> If you could try it out for a while (I don't use global myself other >> than testing a little) and let me know how it goes, that would be >> great. If it works out, I'll check it in. >> >> Thanks >> Eric >> >> *** semanticdb-global.el.~1.5.~ 2009-01-13 19:32:14.000000000 -0500 >> --- semanticdb-global.el 2009-04-18 08:24:09.000000000 -0400 >> *************** >> *** 148,154 **** >> (result (semantic-symref-find-tags-by-name name 'project)) >> ) >> (when result >> ! (semantic-symref-result-get-tags result t)) >> ))) >> >> (defmethod semanticdb-find-tags-by-name-regexp-method >> --- 148,154 ---- >> (result (semantic-symref-find-tags-by-name name 'project)) >> ) >> (when result >> ! (semantic-symref-result-get-tags result)) >> ))) >> >> (defmethod semanticdb-find-tags-by-name-regexp-method >> *************** >> *** 162,168 **** >> (result (semantic-symref-find-tags-by-regexp regex 'project)) >> ) >> (when result >> ! (semantic-symref-result-get-tags result t)) >> ))) >> >> (defmethod semanticdb-find-tags-for-completion-method >> --- 162,168 ---- >> (result (semantic-symref-find-tags-by-regexp regex 'project)) >> ) >> (when result >> ! (semantic-symref-result-get-tags result)) >> ))) >> >> (defmethod semanticdb-find-tags-for-completion-method >> >> >> >> >> >>> peng yu <yup...@gm...> seems to think that: >> >--00163646cef43b665a0467d1adc0 >> >Content-Type: multipart/alternative; >> boundary=00163646cef43b66510467d1adbe >> > >> >--00163646cef43b66510467d1adbe >> >Content-Type: text/plain; charset=ISO-8859-1 >> >Content-Transfer-Encoding: 7bit >> > >> >Sorry, I send the mail twice, because I the first time I forgot add the >> >attach. >> > >> >2009/4/18 peng yu <yup...@gm...> >> > >> >> Hi, Eric. >> >> >> >> 2009/4/17 Eric M. Ludlam <er...@si...> >> >> >> >>> It will certainly open some files, but in general it closes them again >> >>> as needed. >> >>> >> >>> A good trick is to use: >> >> >> >> >> >>> M-x debug-on-entry RET find-file-noselect RET >> >>> >> >>> Then run your experiment. It will stop on every file load. Press "c" >> >>> to continue. It will stop twice per call. Once per entry, and once >> >>> per exit. >> >>> >> >>> Make note of what is on the stack when the offending file goes by. >> >>> That will help direct us to the right thing. If it is related to >> >>> Global and the symref code, we could experiment with letting tags out >> >>> that have line numbers instead of character positions. >> >>> >> >> >> >> >> >On the test code I sent to you, I type M-x debug-on-entry RET >> >find-file-noselect RET >> > >> > Then I type M-x semantic-analyze-current-context, the backtrace stoped >> >twice total. I saved the *Backtrace* buffer contents and send to you. I >> hope >> >you can help me check them, I don't know what is the important thing in >> the >> >*Backrace* buffer. >> > >> >Best Regard. >> >YuPeng >> > >> [ ... ] >> > > |