User forums > Nightly builds

The 15 August 2011 build (7385) is out.

<< < (8/13) > >>

Jenna:
I do not (yet) get lockups with the debugger-branch from my repo (7386), nor with actual trunk.

Both 64-bit linux and build against wx 2.8.10, dbg with gcc 4.3 and trunk with 4.6.
It doesn't matter whether I use one parser per workspace or per project.

What are your settings for codecompletion (parser) and symbols browser ?

Folco:
If you talk that about my post (CC crash, described previous page), I use default settings, I don't change anything.

spectre:

--- Quote from: Zanzicas on August 23, 2011, 06:17:20 pm ---Went to <Settings><Editor...><Code Completion><Symbols Browser page> and checked "Disable Symbols Browser" <Ok>.

--- End quote ---

Ah, I see now.
Well, this time, I disabled symbols browser this way, opened a project (a trivial C++ console application) then again tried to open a .c file that does not belong to this project. C::B hung, like it did before. I tried this both with and without ThreadSanitizer being active, the results are the same.

The log of C::B (cb04.log) and a part of ThreadSanitizer report (tsan-cb04-hybrid.log) are in the attached archive. The race detector still reports some possible data races that could be related to code completion component. I cannot say whether they are real races or whether they are related to this very problem.


--- Quote from: jens on August 23, 2011, 06:31:59 pm ---What are your settings for codecompletion (parser) and symbols browser ?

--- End quote ---

Jens, I use the default settings. Except for disabling symbols browser, I have not changed the settings since I have build C::B (rev. 7406, built on Ubuntu 10.04 x86-64) the first time. default.conf from my system is also in the attached archive, just in case (it is a copy of ~/.codeblocks/default.conf, there were no other .conf files there).

oBFusCATed:
spectre: Is C::B usable, when run under the control of ThreadSanitizer?

spectre:

--- Quote from: oBFusCATed on August 24, 2011, 10:22:47 am ---spectre: Is C::B usable, when run under the control of ThreadSanitizer?

--- End quote ---

It is quite slow but still it is usable.

It took several minutes only to load C::B on my machine under the control of ThreadSanitizer. I suppose it is common with heavyweight Valgrind-based tools (the slowdown factor of MemCheck, for example, can be as high as 20x-50x if I am not mistaken).

I think adjusting ThreadSanitizer settings (enabling sampling, for example) along with providing the machine with more RAM (my box has only 1G) could mitigate this performance issue.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version