Developer forums (C::B DEVELOPMENT STRICTLY!) > CodeCompletion redesign
New code completion plugin w/ refactoring
Ptomaine:
--- Quote from: Takeshi Miya on March 21, 2006, 11:12:16 pm ---
--- Quote from: Ptomaine on March 21, 2006, 10:58:51 pm ---Ok, guys.
First, I'm sorry, guys, for my offensive words towards Arto. I'm really sorry. I didn't mean that.
Well, let's imagine my situation.
I've started using CB about a half year ago. I've created a project and been working on it. I update CB bona fide, time to time with the nightly builds. And continue working. An IDE for me is a kind of a tool that helps me to EDIT my source code, BUILD it and DEBUG it in a convenient way. Generally, it must help me to work. When I said that there was no difference between RC2 and the most recent build then I've ment that the new versions hadn't facilitate eighter my EDITing or BUILDing or DEBUGing processes. For me, CB has changed visually for me but its new features didn't reflect on my work. For instance, I use cups to drink and I really do not care how it looks like. The main thing, the cup lets me drink - it is its general usage (it does the work it's supposed to do) :). So, here it is.
Anyway, guys, I love what you are doing. And sure, you want to turn CB into something amazing.
All my strict words were said because I want CB become something amazing and I care it.
I've just been frightened of myself when I wanted to change the IDE I love (CB) to UPP... so, I've seethed up...
--- End quote ---
I understand all the features you might want, but you must understand two things:
-C::B is under heavy development.
-Lack of human resources (time). All of this is being done by few people in spare time and in a free way.
What you can do from your position is: spread the word, so more developers can join the effort, report bugs, send patches, etc.
Particularly on CodeCompletion, right now, the only devs working (coding) on redesign it are TDragon and ddiego (in separate ways).
You might know that C++ parsing is hell, and so is CodeCompletion, along with threading, performance issues, etc.
It's not an easy topic.
--- End quote ---
Agreed (partially).
Takeshi Miya, look what only two men could create in a short time. Can you compare it to CB, please? (http://upp.sourceforge.net)
thomas:
Well, one could argue a lot about UPP, not only the IDE but also the toolkit. There are quite a few things in UPP which give me a cold shiver... but that is far, far off-topic.
Actually this thread relates to the CC plugin. :)
takeshimiya:
--- Quote from: Ptomaine on March 21, 2006, 11:39:45 pm ---Agreed (partially).
Takeshi Miya, look what only two men could create in a short time. Can you compare it to CB, please? (http://upp.sourceforge.net)
--- End quote ---
Sorry, but the only thing that Ultimate++ haves better than C::B is Code Completion/Code Navigation.
But it's only better, it's not the best. Their hand-crafted C++ parser (Assist++) is not the best, ie. one of the limitations is that it does not expand macro definition and scans only project sources (means standard platform headers are not scanned).
I've researched some months ago on full c++ parsers, and the only ones I have hopes for C::B are:
-CodeStore (currently developed by ddiego, using ANTLR++ and CTags, along with SQLite).
-TDragon's parser (current status unknown, but I think he's still developing it according a few mails with him).
-Eranif's parser (based on CTags and SQLite, already developed, would need porting to GCC and adaptation as a C::B plugin).
Regards,
Takeshi Miya
TDragon:
--- Quote from: Takeshi Miya on March 22, 2006, 12:07:45 am ----TDragon's parser (current status unknown, but I think he's still developing it according a few mails with him).
--- End quote ---
While I'm thinking about it, this seems an appropriate place for a quick update on that. :)
Yes, I'm still working on that as much as I can. I've currently retreated to considering everything that would be involved in writing a full incremental parser for C and C++, but as that would take many moons to reach a useful state I've also got a quick-fix on the burner that would be ready in a shorter time frame.
I also think that if a list were made of people's peeves with the current state of code completion, it would boil down to a majority of relatively easy fixes, plus
- templates and namespaces support, and
- system headers support,
which two could then be left for a later iteration.
Game_Ender:
Full linux/threading stability is also a must.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version