Code::Blocks Forums

User forums => Nightly builds => Topic started by: killerbot on August 23, 2006, 08:28:34 pm

Title: The 23 august 2006 build is out.
Post by: killerbot on August 23, 2006, 08:28:34 pm
Get quick announcements through the RSS feed http://www.codeblocks.org/nightly/CodeBlock_RSS.xml

A link to the unicode windows wxWidget dll for Code::Blocks : http://prdownload.berlios.de/codeblocks/wxmsw26u_gcc_cb_wx2.6.3p2.7z

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

For support of ansi builds, a link to the ansi windows wxWidget dll for Code::Blocks : http://prdownload.berlios.de/codeblocks/wxmsw26_gcc_cb_wx2.6.3p2.7z

The 23 August 2006 build is out.
  - Windows : http://prdownload.berlios.de/codeblocks/CB_20060823_rev2891_win32.7z
  - Linux :
         http://prdownload.berlios.de/codeblocks/CB_20060823_rev2891_Ubuntu6.06.deb
         http://prdownload.berlios.de/codeblocks/CB_20060823_rev2891_fc4+5.rpm


Resolved Fixed:


Regressions/Confirmed/Annoying/Common bugs:


Title: Re: The 23 august 2006 build is out.
Post by: sethjackson on August 23, 2006, 09:11:21 pm
Windows ANSI build is here (http://rapidshare.de/files/30497545/output.7z.html).
Title: Re: The 23 august 2006 build is out.
Post by: MoonKid on August 24, 2006, 07:27:25 am
The fix works just fine. Thanks!

Thanks a lot for the virtual foloder suport. I need it so much.
But something is not right with it.

If I let the tree "display as on disk" the folders are shown correctly with all the files in it. (btw: the virtual folders are not on disc, of course!).
If I disable "display as on disk" the virtual folders are empty and the in-folder-files are in the root of the project-tree displayed with workspacepath+name (for example VirtualFolder\\File.cpp).
Title: Re: The 23 august 2006 build is out.
Post by: Outis on August 24, 2006, 10:41:08 am
Code::Blocks is really great work!  :D
There is just a litte thing I don't like at code-completion: Coding a new header file, I always get the note "no matches", which is, of course, totaly useless in this case. I know that I'm writing new code, code-completion doesn't. Thus could you add a check box to disable the "no matches" - notification, please?
And code-completion doesn't work if there's a this pointer in front of a variable (this->foobar).
Title: Re: The 23 august 2006 build is out.
Post by: MortenMacFly on August 24, 2006, 10:56:13 am
Quote from: Outis on August 24, 2006, 10:41:08 am
And code-completion doesn't work if there's a this pointer in front of a variable (this->foobar).
A bit of topic, but anyway: I never understood why it makes sense to add the this pointer in front. The only reason I see is if one has a local variable with the same name. But this would IMHO be a bad design. Maybe someone can enlighten me?!
Title: Re: The 23 august 2006 build is out.
Post by: Tulio on August 24, 2006, 10:59:53 am
Quote from: Outis on August 24, 2006, 10:41:08 am
Thus could you add a check box to disable the "no matches" - notification, please?

exactly - "no matches" is nervy.
Title: Re: The 23 august 2006 build is out.
Post by: phlox81 on August 24, 2006, 11:29:18 am
Maybe it would be helpful if you could define some more options for the codecompletion,
since every User has its own feeling of how-it-shouldbe.
For myself, I hate when it replaces the typed with the selection, when entering . or -> or ;
I'd like to have replacements only, when hitting enter.
Also annoying is, when you have a vector or an array, you type by open box [] and you get 89.
But thats all my personal flavor, someone else might think totally different, so
to be able to set options for this would be really cool.

phlox
Title: Re: The 23 august 2006 build is out.
Post by: killerbot on August 24, 2006, 11:41:30 am
Quote from: MortenMacFly on August 24, 2006, 10:56:13 am
Quote from: Outis on August 24, 2006, 10:41:08 am
And code-completion doesn't work if there's a this pointer in front of a variable (this->foobar).
A bit of topic, but anyway: I never understood why it makes sense to add the this pointer in front. The only reason I see is if one has a local variable with the same name. But this would IMHO be a bad design. Maybe someone can enlighten me?!

It is totally bad design. One just does not do that.
A common used approach to avoid collision with an 'equal' named local is :
 m_MyMember
 _MyMember
 MyMember_
Title: Re: The 23 august 2006 build is out.
Post by: Phoenix on August 24, 2006, 12:09:32 pm
And back to important things. Class browser is working even worst then in nightly from 22/08. Right now each time you select "reparse" it just closes down C::B with no crash raport.
Title: Problem with virtual folders
Post by: jpaterso on August 24, 2006, 12:38:38 pm
When I create a virtual folder directly under my project, (so I end up with 3 folders, 'virtual', 'sources' and 'headers'), and only put headers in it, then when I restart C::B it puts that folder under the 'headers' folder. Would it be possible to keep our folders where we left them?
Title: Re: The 23 august 2006 build is out.
Post by: mdelfede on August 24, 2006, 01:42:18 pm
Quote from: MortenMacFly on August 24, 2006, 10:56:13 am
Quote from: Outis on August 24, 2006, 10:41:08 am
And code-completion doesn't work if there's a this pointer in front of a variable (this->foobar).
A bit of topic, but anyway: I never understood why it makes sense to add the this pointer in front. The only reason I see is if one has a local variable with the same name. But this would IMHO be a bad design. Maybe someone can enlighten me?!

In template code I found that 'this' is required, if not gcc gives an error. Don't know why.

Bad one :
Code
template<class BClass, class ImplClass> class DerClass : public BClass
{
  public:

    DerClass() : BClass() { impl = new ImplClass();}                       <--- ERROR, 'impl' NOT FOUND
    ImplClass *getImpl() { return dynamic_cast<ImplClass *>(this->impl);}

};

Good one :
Code
template<class BClass, class ImplClass> class DerClass : public BClass
{
  public:

    DerClass() : BClass() { this->impl = new ImplClass();}
    ImplClass *getImpl() { return dynamic_cast<ImplClass *>(this->impl);}

};

And somebody uses it in constructors too...
Code
X:X(int a, b, c)
{
  this->a = a;
  this->b = b;
.....
I find that one awful, but...

Ciao

Max
Title: Re: The 23 august 2006 build is out.
Post by: marfig on August 24, 2006, 02:08:05 pm
OT: The compiler doesn't look for base classes unqualified identifiers that are depending on a template parameter. So this-> becomes a necessity under those situations. The problem is that if you don't, then errors could only be caught after the template was instantiated. By using this-> you can give the compiler the opportunity to catch errors much sooner, during the parsing phase.

There is also of course always the problem of code readability. With the right conditions in complexity, it is quiet possible the use of this-> will make the code more readable even without templates.
Title: Re: The 23 august 2006 build is out.
Post by: Pecan on August 24, 2006, 02:59:29 pm
Quote from: MortenMacFly on August 24, 2006, 10:56:13 am
Quote from: Outis on August 24, 2006, 10:41:08 am
And code-completion doesn't work if there's a this pointer in front of a variable (this->foobar).
A bit of topic, but anyway: I never understood why it makes sense to add the this pointer in front. The only reason I see is if one has a local variable with the same name. But this would IMHO be a bad design. Maybe someone can enlighten me?!

Meyers has an example in the first book of the set showing a necessary use of "this" in multiple inheritance.
Title: Re: The 23 august 2006 build is out.
Post by: MoonKid on August 24, 2006, 07:17:56 pm
Another thing about the virtual folders.
It is not possible to drag&drop virtual folders. For example I want to drag an existing folder as a sub-folder to another one.
Title: Re: The 23 august 2006 build is out.
Post by: mandrav on August 25, 2006, 12:32:12 pm
Quote from: MoonKid on August 24, 2006, 07:17:56 pm
Another thing about the virtual folders.
It is not possible to drag&drop virtual folders. For example I want to drag an existing folder as a sub-folder to another one.

Not implemented yet.
Title: Re: The 23 august 2006 build is out.
Post by: mandrav on August 25, 2006, 02:34:01 pm
Quote from: mandrav on August 25, 2006, 12:32:12 pm
Quote from: MoonKid on August 24, 2006, 07:17:56 pm
Another thing about the virtual folders.
It is not possible to drag&drop virtual folders. For example I want to drag an existing folder as a sub-folder to another one.

Not implemented yet.

Implemented now :).
Title: Re: The 23 august 2006 build is out.
Post by: nix_BB on August 25, 2006, 06:19:57 pm
that was quick!