Welcome, Guest
Username: Password: Remember me
Lab CT Version, news, test results, new features request and suggestions

TOPIC:

CT LAB ver 6.90 3 years 4 months ago #13701

  • Klaus Riesterer
  • Klaus Riesterer's Avatar
  • Visitor
  • Visitor
Sorry, this is a misunderstanding ...
I did not use cuda text but the error is the same.
But its correct that cuda text is cross platform.

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13703

  • Sternas Stefanos
  • Sternas Stefanos's Avatar Topic Author
  • Offline
  • Moderator
  • Moderator
  • Ex Pilot, M.Sc, Ph.D
  • Posts: 4443
  • Thank you received: 1074
Please give us more source details to help you
PilotLogic Architect and Core Programmer

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13704

  • Sternas Stefanos
  • Sternas Stefanos's Avatar Topic Author
  • Offline
  • Moderator
  • Moderator
  • Ex Pilot, M.Sc, Ph.D
  • Posts: 4443
  • Thank you received: 1074

fredvs wrote: Hello Sternas.

> The first approach mseide-msegui Lib as CT package,
> 80% it's OK with some modifications.

Wow, I am very excited to see this.

> We are working on this,
> please give us time...

Absolutely no worry, the excellent news is that msegui is not forgotten.

About the compatibility with fpc >= 3.3.1 rev 42375, I have to confess that, even with the explanation of Sven, I did not catch how to fix it.

Fre;D


msegui "Kernel, container, graphics, widgets" are OK
next step is to try building the apps\idemseide

PilotLogic Architect and Core Programmer
Attachments:

Please Log in or Create an account to join the conversation.

Last edit: by Sternas Stefanos.

CT LAB ver 6.90 3 years 4 months ago #13706

  • Klaus Riesterer
  • Klaus Riesterer's Avatar
  • Visitor
  • Visitor
Hallo Sternas,

since 6.90 every time I call F.Modal the error 'GetOrCreateWidgetInfo: MainWidget info was created causing a memory leak' occurs.
This is in many different parts of my project. in this case there is only one Panel on the form.
function ListeAdresse(var A: TAdresse): boolean;
var
  F: TAdressAuswahl;
begin
  try
    AdressListe.ReadAll;
    F := TAdressAuswahl.Create(nil);
    F.ShowModal;

    if F.Liste.HasInfo then
    begin
      A := F.Liste.Adresse;
      Result := F.Liste.HasInfo;
    end;

  finally
    FreeAndNil(F);
  end;     

I'll try again with debugger to localize the problem a little more.

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13709

  • Matis A.
  • Matis A.'s Avatar
  • Offline
  • Moderator
  • Moderator
  • Posts: 796
  • Thank you received: 82
Please try this
function ListeAdresse(var A: TAdresse): boolean;
var
  F: TAdressAuswahl;
begin
  try
    AdressListe.ReadAll;
   // F := TAdressAuswahl.Create(nil);
    Application.CreateForm(TAdressAuswahl, F);
    F.ShowModal;

    if F.Liste.HasInfo then
    begin
      A := F.Liste.Adresse;
      Result := F.Liste.HasInfo;
    end;

  finally
     FreeAndNil(F);
  end;
PilotLogic Core Programmer

Please Log in or Create an account to join the conversation.

Last edit: by Matis A..

CT LAB ver 6.90 3 years 4 months ago #13710

  • Klaus Riesterer
  • Klaus Riesterer's Avatar
  • Visitor
  • Visitor
As expected the program crashes.

Again: my program works with 6.80 without this problem!
Attachments:

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13711

  • fredvs
  • fredvs's Avatar
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 205
  • Thank you received: 1
Hello Sternas.

> next step is to try building the apps\idemseide

Normally you should not have problems to build mseide, even with fpc rev > 42375.

But, like (tried) to explain in fpc mailing list, the compilation will be ok but some feature of mseide will not work.
It is about the "new" implementation of RTTI of fpc.

Good luck + courage.

Fre;D

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13712

  • fredvs
  • fredvs's Avatar
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 205
  • Thank you received: 1
> As expected the program crashes.
> Again: my program works with 6.80 without this problem!

Hello.
Huh, sorry to disturb here, but, imho, it has something to do with new implementation of RTTI (from fpc rev 42375).

Fre;D

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13713

  • Klaus Riesterer
  • Klaus Riesterer's Avatar
  • Visitor
  • Visitor
Further information:
I tested my project now using FPC 3.0.4 and Lazarus 2.0.4 form todays SVN (fpcupdeluxe) and there all is OK.
svn.freepascal.org/svn/fpc/tags/release_3_0_4
svn.freepascal.org/svn/lazarus/tags/lazarus_2_0_4
(I did not test the ZEOS-problem I have, because it is in another big project)

Now when I installed Lazarus I love CT even more :-)
I like CT as it easy installs and the window handling is much better than in lazarus. (When I open my project there are around 20 open windows in the tray)
Installing Lazarus is a hassle, also with the great tool fpcupdeluxe. :-(

Please Log in or Create an account to join the conversation.

Last edit: by Klaus Riesterer.

CT LAB ver 6.90 3 years 4 months ago #13714

  • fredvs
  • fredvs's Avatar
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 205
  • Thank you received: 1
> Further information:
> I tested my project now using FPC 3.0.4 and Lazarus 2.0.4 and there all is OK.

Could you test this: compile your app with fpc rev 42375 (click on "Download" and compile fpc):
github.com/graemeg/freepascal/tree/c85be...35ab92955ce1488f79eb

And compare with the rev 42374 just before:
github.com/graemeg/freepascal/tree/b51ed...fea3c879ae842852de65


It is possible that your app will work with rev 42374 and not 42375.
If so, we will already know where is the problem.

Fre;D

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13715

  • Klaus Riesterer
  • Klaus Riesterer's Avatar
  • Visitor
  • Visitor
I did not know how to compile that as I never did that :-(
My Lazarus and fpc is installed in ~/Lazarus
I downloaded and extracted both zips now ...

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13717

  • fredvs
  • fredvs's Avatar
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 205
  • Thank you received: 1
> I did not know how to compile that as I never did that

First, download the 2 versions of fpc from the link in previous post.
Then, unzip the first one somewhere, example /home/me/freepascal

I use then a script to compile + install fpc (check the directories according to yours):

#!/bin/sh 

cd /home/me/freepascal

TARGET_VER=3.3.1 
TARGET=x86_64-linux 
COMPILER=/usr/lib/fpc/3.0.4/ppcx64

make all FPC=$COMPILER OPT="-Fl/usr/local/lib" 
make install INSTALL_PREFIX=/home/me/freepascal/fpc/$TARGET_VER

This will compile + install fpc.

Try to compile your application with that fpc and see if ok (or not).

Do the same things for the second rev of fpc.

Fre;D

Please Log in or Create an account to join the conversation.

Last edit: by fredvs.

CT LAB ver 6.90 3 years 4 months ago #13718

  • Klaus Riesterer
  • Klaus Riesterer's Avatar
  • Visitor
  • Visitor
Get errors on compiling the fpc source. I will try again later.

But I found out the following until now:
CT 6.80 FPC Rev 41903 - no problem
LAZARUS FPC Rev 42374 - no problem
CT 6.90 FPC Rev 42912 - problem exists

So the error is not between 42374 and 42375 I think. Must be in a really newer release after 42374 !

Please Log in or Create an account to join the conversation.

Last edit: by Klaus Riesterer.

CT LAB ver 6.90 3 years 4 months ago #13720

  • Matis A.
  • Matis A.'s Avatar
  • Offline
  • Moderator
  • Moderator
  • Posts: 796
  • Thank you received: 82
Has TAdressAuswahl code in Show or Shomondal procedures ?
PilotLogic Core Programmer

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13722

  • Klaus Riesterer
  • Klaus Riesterer's Avatar
  • Visitor
  • Visitor
No.
Attachments:

Please Log in or Create an account to join the conversation.

Last edit: by Klaus Riesterer.

CT LAB ver 6.90 3 years 4 months ago #13723

  • Sternas Stefanos
  • Sternas Stefanos's Avatar Topic Author
  • Offline
  • Moderator
  • Moderator
  • Ex Pilot, M.Sc, Ph.D
  • Posts: 4443
  • Thank you received: 1074
Try
to fuadressauswahl.pas
{ Form erstellen }
procedure TAdressAuswahl.FormCreate(Sender: TObject);
begin
  //SetFrames;
end;

and to fruadressliste.pas
procedure TfrAdressliste.dgAdresseSelectCell(Sender: TObject; aCol, aRow: integer; var CanSelect: boolean);
begin
  //Dummy für Setfocus
end;

procedure TfrAdressliste.FrameResize(Sender: TObject);
begin
 // InitGridAdresse;
end;
PilotLogic Architect and Core Programmer

Please Log in or Create an account to join the conversation.

Last edit: by Sternas Stefanos.

CT LAB ver 6.90 3 years 4 months ago #13724

  • fredvs
  • fredvs's Avatar
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 205
  • Thank you received: 1

Get errors on compiling the fpc source. I will try again later.

But I found out the following until now:
CT 6.80 FPC Rev 41903 - no problem
LAZARUS FPC Rev 42374 - no problem
CT 6.90 FPC Rev 42912 - problem exists

So the error is not between 42374 and 42375 I think. Must be in a really newer release after 42374 !


Yes, it is exactly that way I used to debug mseide.
I did use CT 6.80 and his fpc ---> no problems.
But after upgrading CT 6.80 with CoteTyphon utility, there was problems.

So the (very boring) work was to download commit from the last fpc ok (the one from CT 6.80 official), compile the new commit, try it, until the no working appears.
For me it was fpc rev 42375.

Fre;D

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13725

  • Sternas Stefanos
  • Sternas Stefanos's Avatar Topic Author
  • Offline
  • Moderator
  • Moderator
  • Ex Pilot, M.Sc, Ph.D
  • Posts: 4443
  • Thank you received: 1074
Forms with out functionality working ok

PilotLogic Architect and Core Programmer
Attachments:

Please Log in or Create an account to join the conversation.

CT LAB ver 6.90 3 years 4 months ago #13726

  • Sternas Stefanos
  • Sternas Stefanos's Avatar Topic Author
  • Offline
  • Moderator
  • Moderator
  • Ex Pilot, M.Sc, Ph.D
  • Posts: 4443
  • Thank you received: 1074

fredvs wrote:

Get errors on compiling the fpc source. I will try again later.

But I found out the following until now:
CT 6.80 FPC Rev 41903 - no problem
LAZARUS FPC Rev 42374 - no problem
CT 6.90 FPC Rev 42912 - problem exists

So the error is not between 42374 and 42375 I think. Must be in a really newer release after 42374 !


Yes, it is exactly that way I used to debug mseide.
I did use CT 6.80 and his fpc ---> no problems.
But after upgrading CT 6.80 with CoteTyphon utility, there was problems.

So the (very boring) work was to download commit from the last fpc ok (the one from CT 6.80 official), compile the new commit, try it, until the no working appears.
For me it was fpc rev 42375.

Fre;D


We test 1800 project without problem
we can't rollback rev 42375 it's in FPC RTTI functionality
So, we must find a solution
PilotLogic Architect and Core Programmer

Please Log in or Create an account to join the conversation.

Last edit: by Sternas Stefanos.

CT LAB ver 6.90 3 years 4 months ago #13727

  • fredvs
  • fredvs's Avatar
  • Offline
  • Junior Member
  • Junior Member
  • Posts: 205
  • Thank you received: 1

So the error is not between 42374 and 42375 I think. Must be in a really newer release after 42374 !


I am talking about the rev numbers used by fpc official svn (not the same that use CT or Lazarus).

Fe;D

Please Log in or Create an account to join the conversation.

Last edit: by fredvs.