The 17 July 2007 build (4273) is out.

[killerbot]

Get quick announcements through the RSS feed http://www.codeblocks.org/nightly/CodeBlock_RSS.xml

Before you use a nightly make sure you understand how it works.

A link to the unicode windows wxWidget dll for Code::Blocks : http://prdownload.berlios.de/codeblocks/wxmsw28u_gcc_cb_wx284.7z

For those who might need this one (when no MingW installed on your system) : the mingw10m.dll : http://prdownload.berlios.de/codeblocks/mingwm10.7z

The 17 July 2007 build is out.
  - Windows :
   http://prdownload.berlios.de/codeblocks/CB_20070717_rev4273_win32.7z
  - Linux :
   http://prdownload.berlios.de/codeblocks/CB_20070717_rev4273_Ubuntu6.10+7.04_wx2.8.4.deb (not yet)
   http://prdownload.berlios.de/codeblocks/CB_20070717_rev4273_Debian4.0_wx2.8.4.deb (not yet)
   http://prdownload.berlios.de/codeblocks/CB_20070717_rev4273_suse100-102.wx28.i586.rpm (not yet)
   http://prdownload.berlios.de/codeblocks/CB_20070717_rev4273_fc4+5.i586.rpm (not yet)


Resolved Fixed:

• Debian dir updated to create the following packages:
  - libcodeblocks0 (codeblocks shared lib)
  - codeblocks (main app and core plugins)
  - codeblocks-contrib (contrib plugins)
  - codeblocks-dbg (debugging symbols)
  - codeblocks-devel (SDK headers)
  - libwxsmithlib0 (wxSmith shared lib)
  - libwxsmithlib0-devel (wxSmith development shared lib)
• Added rudimentary man pages for cb_share_config and codesnippets
• Resource zip-files are now automatically generated when their dependencies change (no more need for "make clean-zipfiles").
• update-revision.sh now updates debian/changelog too so generated packages will contain correct version info
• Some code-cleanup and improvements in OpenWatcom option generation code
• Added experimental support of creating Linux executables on Windows using OpenWatcom. Please note that an additional **Library directory** has to be added manually

Regressions/Confirmed/Annoying/Common bugs:

• toolbar-images-not-changing-state (is a wx problem/Win XP problem)

[Grom]

I wont to have RC3 :cry:

[kurapix]

Nice ... I feel that RC3 is not that far away (didn't mandrav said he was keeping the new debian dir? ).

Kurapix

[indigo0086]

Are there extra steps I need to do to get symbols for Boost libraries to show.  I can use them, but even when I try to recal variables I just declared they don't show up.

[drrngrvy]

I took a short break from reinstalling the debian packages of the nightlies for the last few weeks. Now I'm back I can't get them working: I'm told "Dependency is not satisfiable: libwxbase2.8-0" despite the fact that I've got libwxbase2.8-0 and libwxbase2.8-dev installed (I have version 2.8.1.1-0ubuntu4). Has nobody else had this problem?

[darthdespotism]

I think CB requires 2.8.4 if you want a 2.8.*

[cstudent]

I took a short break from reinstalling the debian packages of the nightlies for the last few weeks. Now I'm back I can't get them working: I'm told "Dependency is not satisfiable: libwxbase2.8-0" despite the fact that I've got libwxbase2.8-0 and libwxbase2.8-dev installed (I have version 2.8.1.1-0ubuntu4). Has nobody else had this problem?

The deb requires you to install wx2.8.4 using the wxWidgets repository. Read the wiki page on how to update everything.

[rickg22]

I've got a problem with build 4278...

The code completion hints are back! But everytime i hover empty space, I get this annoying hint:

(const char*, int, int)
(const char*)
(int, const char*)
(const char*, const char*)
(int)
(const char*)

And it's making me mad!!!  :evil: Any idea of how to fix it?

[kisoft]

I've got a problem with build 4278...

The code completion hints are back! But everytime i hover empty space, I get this annoying hint:

(const char*, int, int)
(const char*)
(int, const char*)
(const char*, const char*)
(int)
(const char*)

And it's making me mad!!!  :evil: Any idea of how to fix it?

I see another hint (see attach). Really annoying.


[attachment deleted by admin]

[raph]

I've got a problem with build 4278...

The code completion hints are back! But everytime i hover empty space, I get this annoying hint:

(const char*, int, int)
(const char*)
(int, const char*)
(const char*, const char*)
(int)
(const char*)

Exactly the same here.
I temporarily disabled it, until it is fixed.

Some other issues:

• If a hint is shown and you alt+tab to another app, the hint doesn't hide, but will be visible all the time. Sometimes I can't even get rid of it
• If a hint is shown and I switch to a helpfile, codeblocks steals focus and comes to top (haven't experienced this behavior with other apps, yet)

[attachment deleted by admin]

[dje]

Hi !

The same for me when I type

Code Select Expand

if (

Dje

[Wolf]

Hello,
I simply opened my project, and as soon as my mouse was over a space, it showed the annoying hint everyone seems to be getting. I didn't wait at all on a space, it just displayed it right away. And now it is hiding part of this page. =P I even get two of them, but the second one disappears after a while. (yay! still turned it off)

Also, I noticed something strange. I don't think C::B parses the files as it should for code completion. Here's the folder architecture:
include
  SDL
    SDL.h
    SDL_video.h
    ...

#include <SDL/SDL.h>

What I see:

• variables/constants + functions + preprocessor symbols[SDL.h]
• preprocessor symbols[files included by SDL.h]

What I don't but should see:

• variables/constants + functions [files included by SDL.h]

It seems like it is recursive only for preprocessor symbols.
Hope I gave enough details.

-- Wolf --

[rickg22]

so, how do you disable that popup hint?

[MortenMacFly]

so, how do you disable that popup hint?

Settings -> Editor -> Code Completion -> Tab "Code completion" -> deactivate "Display info when hovering mouse over a token".
With regards, Morten.
Ps.: BTW: now that I have compiled a more recent version of C::B this stroke me, too... quite annoying. :-(

[foomaster]

Hello,
I simply opened my project, and as soon as my mouse was over a space, it showed the annoying hint everyone seems to be getting. I didn't wait at all on a space, it just displayed it right away. And now it is hiding part of this page. =P I even get two of them, but the second one disappears after a while. (yay! still turned it off)

Also, I noticed something strange. I don't think C::B parses the files as it should for code completion. Here's the folder architecture:
include
  SDL
    SDL.h
    SDL_video.h
    ...

#include <SDL/SDL.h>

What I see:

• variables/constants + functions + preprocessor symbols[SDL.h]
• preprocessor symbols[files included by SDL.h]

What I don't but should see:

• variables/constants + functions [files included by SDL.h]

It seems like it is recursive only for preprocessor symbols.
Hope I gave enough details.

-- Wolf --

There is something even stranger about the parser. I've been developing a small project (Hangman) + a small framework on top of SDL + STL for some time now, and I've been using Code::Blocks, because I do intend to make my program available on gcc-enabled os-es. Today I was using the build from 17th July, not 18th, so I'll speak about it. I've risen the question of cc's search depth and was ensured that it was recursive and only the "finding of files" was the problem. The parser has truly evolved this week, I'm impressed. But there are some illogical things about it. For example:
1) I use MSYS + MinGW as a compiler set (mingw being placed in the recommended %msys_root%/mingw dir)
2) My SDL headers are in the usual directory for a windows install (%msys_root%/include/SDL)
The bug (?)

Code Select Expand

...
#include <SDL/SDL.h>
...
namespace Hangman {
...
   SDL_Surface* example = NULL;
... // load / create the surface
   Uint32 encodedColour = SDL_MapRGB(example->format, 0, 0xff, 0xff); // <===
...
}
Here we have all kinds of suggestions, but not format. Ok, I'm used to Kate's approach to "code completion", so it's not a problem to remember things. However, I'm failing to understand why the hell it doesn't have the whole list of suggestions. If it were a #define, it wouldn't be parsed at all (furthermore, I don't see it realized as a #define...). There are suggestions, so it IS parsed. Then, why not the whole list of members?

Bug (?) #2
The default "function" completion. Whenever (I don't know if it actually reproduces itself every single time, but it is reproducible this way) an opening bracket "(" is typed and there is no default argument list (that is, there is no function declared before that moment with the given name), a "default" parameter list is evoked. It is identical to to tool-tip "bug" list. That is why I think they are closely related. The problem is: this is truly annoying. If I type "...main(", the list is evoked. If I type "if (", "while (", or any other keyword (which makes me think the parser can't fully tell a keyword apart from identifier, let alone function/keyword) the same default function argument list is evoked.

Bug (?) #3
Local variables aren't truly "local". I mean, sometimes I have a local member-function variable called (e.g.) result. As far as I'm aware, this is neither c++, nor c keyword. Bah, this doesn't matter, as this occurs with any local variable. So, I have my variable in one function and then I use the same name in another (this happens even when they are in different files 0_o), it is auto-completed (in the list of suggestions). This gets even more interesting: scope seems to be a fierce adversary. I use heavily SDL_Color. This is a structure with four 8-bit integers, named r, g, b and unused. Well, I noticed that there was something strange with "unused", so I typed it without "m_clr." ... and guess what? It was in the list of suggested items. Truly amazing.

Well, I can't think of more weird things about Code::Blocks, but, hell, it's late and I'm supposed to be getting early this morning!

Cheers, and keep up the good work!