Re: [fpc-devel] Free Pascal Extender

2010-10-30 Thread Alexander Klenin
On Sat, Oct 30, 2010 at 16:53, Hans-Peter Diettrich
drdiettri...@aol.com wrote:
 The SourceForge fpcext project http://sourceforge.net/p/fpcext/home/ shall
 allow to customize and extend the FPC compiler, with e.g. alternative
 components and languages.

Are you sure you do not want to use Github?

-- 
Alexander S. Klenin
___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel


Re: [fpc-devel] Free Pascal Extender

2010-10-30 Thread Graeme Geldenhuys
On 30 October 2010 09:59, Alexander Klenin  wrote:

 Are you sure you do not want to use Github?

There is nothing wrong with SourceForge.net't git setup, it works just
as good as GitHub. The benefit of SourceForge is that they give you
plenty of other project related tools too. eg: a location for a
website, various bug trackers, wikis, mailing lists etc etc... So for
overall project management, SourceForge is a much better choice.

-- 
Regards,
  - Graeme -


___
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net:8080/fpgui/
___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel


Re: [fpc-devel] need to call FpFlock to unlock file in FileClose?

2010-10-30 Thread Michael Van Canneyt



On Sat, 30 Oct 2010, Vinzent Höfler wrote:

On Sat, 30 Oct 2010 00:18:15 +0200, Seth Grover sethdgro...@gmail.com 
wrote:



My question is this: if FileOpen automagically handles the file
locking through a call to DoFileLocking, wouldn't it be more correct
for FileClose handle unlocking as well?


It should, if it's necessary.

Interestingly

man 2 flock says:

|Furthermore, the lock is released either by an explicit LOCK_UN operation on
|any of these duplicate descriptors, or when all such descriptors have been
|closed.

So I would expect the lock to disappear when the file is closed. Nitpicking
people might say, this only applies to duplicated descriptors, but then
what's the point?

So, actually I'm inclined to consider that a bug in the underlying OS.


I tend to agree with this. The reason the explicit unlock is not done is that 
the close should do it implicitly (as per the man page).


But are you sure the first close() operation actually succeeded ?

Michael.___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel


Re: [fpc-devel] Free Pascal Extender

2010-10-30 Thread Hans-Peter Diettrich

Graeme Geldenhuys schrieb:

On 30 October 2010 09:59, Alexander Klenin  wrote:

Are you sure you do not want to use Github?


There is nothing wrong with SourceForge.net't git setup, it works just
as good as GitHub. The benefit of SourceForge is that they give you
plenty of other project related tools too. eg: a location for a
website, various bug trackers, wikis, mailing lists etc etc... So for
overall project management, SourceForge is a much better choice.


Just encountered a problem: when I add FPDoc documentation, currently 
residing in compiler/docs/, the content should be available in all 
branches, updates should not be affected by checking out other branches.


Questions to the gurus:

How can I add this directory to the git repository, so that it is 
available as a self-contained project/branch?


Since most of the (current) documentation applies to FPC in general, 
where should such documentation reside, in the (SVN) directory tree? How 
to separate general docs from branch-related docs?


DoDi

___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel


Re: [fpc-devel] Free Pascal Extender

2010-10-30 Thread Graeme Geldenhuys
On 30 October 2010 15:43, Hans-Peter Diettrich wrote:
 How can I add this directory to the git repository, so that it is available
 as a self-contained project/branch?

SourceForge.net allows unlimited git repositories per project. Each
repository is stand-alone, but under the same sourceforge project.
Not sure if this is what you want, but there is another option. I
think it's called git modules, where you can have multiple
repository working together in one hierarchy. I have no idea how to
set this up though, I have never used that feature. I know KDE project
does. Just read the git docs, they should explain it there.


-- 
Regards,
  - Graeme -


___
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net:8080/fpgui/
___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel


[fpc-devel] Problem with cascading Destructor in fpc 2.4

2010-10-30 Thread Martin Krüger
Hi

I introduced an boolean-option to the destructor of a class. I
implemented  a wrapper without argument  for reason of
backward-compability.

This wrapper calls the real destructor with default argument.

Everything worked fine until fpc 2.4  (sample programm below)
When running it compiled with fpc 2.4 i get an Runtime error 216
(segmentation fault).
I could reproduce the problem with the latest svn head.

Without further examination i suggest: The outer destructor tries to
free the memory the inner destructor already released.

So my question:
Is it a bug or am i doing something i should not?
The compiler gives me am note:

test.pas(8,3) Note: Class should have one destructor only


-- 
Best regards,
  Martin






program cascading_destructor;

type

test=class
  constructor create();
  destructor  destroy();overload;
  destructor  destroy(arg:boolean);overload;
end;


constructor test.create();
begin
end;

destructor  test.destroy();
begin
  writeln('calling master destroy');
  destroy(true);
  writeln('master destroy done');
end;

destructor  test.destroy(arg:boolean);
begin
  writeln('destroyed');
end;


var a:test;

begin
  a:=test.create();
  a.destroy;
end.
type

test=class
  constructor create();
  destructor  destroy();overload;
  destructor  destroy(arg:boolean);overload;
end;


constructor test.create();
begin
end;

destructor  test.destroy();
begin
  writeln('calling master destroy');
  destroy(true);
  writeln('master destroy done');
end;

destructor  test.destroy(arg:boolean);
begin
  writeln('destroyed');
end;


var a:test;

begin
  a:=test.create();
  a.destroy;
end.


___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel


Re: [fpc-devel] Problem with cascading Destructor in fpc 2.4

2010-10-30 Thread Luca Olivetti

Al 30/10/10 17:54, En/na Martin Krüger ha escrit:



So my question:
Is it a bug or am i doing something i should not?
The compiler gives me am note:

test.pas(8,3) Note: Class should have one destructor only


If I remember correctly you should never call destroy directly, only 
free (which in turn will call your destructor).


Bye
--
Luca
___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel


Re: [fpc-devel] Problem with cascading Destructor in fpc 2.4

2010-10-30 Thread Michael Van Canneyt



On Sat, 30 Oct 2010, Martin Krüger wrote:


Hi

I introduced an boolean-option to the destructor of a class. I
implemented  a wrapper without argument  for reason of
backward-compability.

This wrapper calls the real destructor with default argument.

Everything worked fine until fpc 2.4  (sample programm below)
When running it compiled with fpc 2.4 i get an Runtime error 216
(segmentation fault).
I could reproduce the problem with the latest svn head.

Without further examination i suggest: The outer destructor tries to
free the memory the inner destructor already released.

So my question:
Is it a bug or am i doing something i should not?
The compiler gives me am note:

test.pas(8,3) Note: Class should have one destructor only


This is your problem.

The destructor must always be called destroy (overriding TObject.Destroy)
and should accept no arguments. If you want to pass an argument to it somehow, 
use a property which you set prior to calling destroy.


Michael.___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel


Re: [fpc-devel] Problem with cascading Destructor in fpc 2.4

2010-10-30 Thread Vinzent Höfler

On Sat, 30 Oct 2010 17:54:13 +0200, Martin Krüger martin.krue...@gmx.com
wrote:


I introduced an boolean-option to the destructor of a class.


Actually, you seem to be violating basic object oriented paradigms here. An
object should always know from its current state how to destruct itself, so
why should a destructor need a parameter?


Vinzent.
___
fpc-devel maillist  -  fpc-devel@lists.freepascal.org
http://lists.freepascal.org/mailman/listinfo/fpc-devel