Hi

If you start Lazarus and wait until every dialog at startup is shown and change then the console (the splashscreen must be shown, too) (only tested under KDE 2) and when I then change back to the Lazarus console after some seconds I get the following SEGFault:

An unhandled exception occurred at 0x080DF606 :
Access violation
0x080DF606 TBRUSH__GETHANDLE, line 134 of ./include/brush.inc 0x00000042
0x00000001
An unhandled exception occurred at 0x08100F2A :
Access violation
0x08100F2A
0x3020200C main, line 42 of lcllinux.pp
0x00000000

An unhandled exception occurred at 0x08105D3C :
Access violation
0x08105D3C
0x2C35313A main, line 42 of lcllinux.pp
0x1667E605 main, line 42 of lcllinux.pp

[TgtkObject.Destroy] WARNING: There are 138 unreleased DCs and 489 unreleased GDIObjects

cu Jens

(possibly

a shutdown) and there are no processes left to kill. Normally you can

ignore

this message.

Ignoring is not possible, because the system hangs, only a reset helps. This message is declareing itself in the normal run of the system that every dynamically linked program I am trying to start is SEGFaulting. If I then do a reboot or halt the error message appears. What can the reason for this message be? Could the reason be a buggy RAM or Cache-RAM?

Jens

>>>>>>>>>>>>>>>>> Urspr�ngliche Nachricht <<<<<<<<<<<<<<<<<<

Am 2/19/01, 9:51:33 AM, schrieb (Marco van de Voort) zum Thema Re: [lazarus] New tests of the IDE:

In the first case: update INIT, in the second case, upgrade the kernel,

in

other cases, I don't know.

I use Kernel 2.4 (SuSE 7.1)

_________________________________________________________________ To unsubscribe: mail with "unsubscribe" as the Subject archives at http://www.miraclec.com/list_archives/lazarus

Do you use

A hand upgraded system dated back to 2.0.3x or even 2.0.2x days?

or

an early 2.2.x kernel? (x <=5 or so)

In the first case: update INIT, in the second case, upgrade the kernel, in other cases, I don't know.

Hi

I have used the weekend to test a bit the IDE and I have found some more not correct working things:

1. The view Units and Forms Dialogs are not working correctly (it is only one unit shown). (the view Forms can not be executed).

2. Open Project menuitem is not working.

3. If I drop non visible controls, like timers, on the form in the formeditor I get a SEGFault.

4. The grabber in the formeditor are not set correctly if you change e.g. between a button and the form. You can not use the grabbers for resizing the controls.

5. The mousepointer changes to a down arrow if I am in the object inspector.

6. The objectinspector is to bit of it is
hidden by the maindialog.

7. Speedbutton for changing between Form and Unit is not working correctly.

8. The file-close menuitem only closes the unit not the corresponding form.

9. A replace-text-dialog is missing for the editor.

10. Can someone remove the executable bits from the lazarus.dci in the CVS?

11. It is now impossible to mark lines or words with the mouse, but dbl-click marking of one word is working.

12. If I select new project only units are generated, not the corresponding form and the generated units are different in the code to the units which are generated over the menu (File-new Form).

13. ListBox: It would be nice if it is possible to scroll with the cursorkeys through the list and select the wanted item with the spacebar.

14. What is with the tooltips, are they disabled again?

15. One time I had that in the editor the cursor was shown at a fixed position but I could write at any place.

16. The CTRL-Space function isn't working if I say open project over the project menu.

17. If I close Lazarus I get the following error:

Gtk-CRITICAL **: file gtkmain.c: line 534 (gtk_main_quit): assertion `main_loops
!= NULL' failed.
An unhandled exception occurred at 0x08104CBC :
Access violation
0x08104CBC
0x0806278C THEFONTSINFOMANAGER__RELEASEFONTSINFO, line 517 of syntextdrawer.
pp
0x08062C52 THEFONTSTOCK__RELEASEFONTSINFO, line 735 of syntextdrawer.pp
0x080629E9 THEFONTSTOCK__DESTROY, line 631 of syntextdrawer.pp 0x08104CBF
0x08062EEC THETEXTDRAWER__DESTROY, line 836 of syntextdrawer.pp 0x08104CBF
0x0806B8A9 TCUSTOMSYNEDIT__DESTROY, line 1123 of synedit.pp 0x080F5E2B
0x080F5DD7

PS:
Does someone of you know what the Errormessage (init: no more processes left in this runlevel) means and what the matter for this message can be???

cu Jens

At home I build Lazarus from the CVS-Source, but if I be somewhere else I use the daily build lazarus.zip from the freepascal-ftp.

<snip - error report>

I just wanted to give a thanks to this kind of error report. This is the kind of testing we need to find all of the kinks in the armor. We realize that this is difficult if you don't get the source and build the IDE from CVS. In a few days we should have the bugs worked out of building nightly binaries of the IDE. Then you can download and test the IDE for us.

Matthias, what types of controls do you mean with non-wincontrol?? I thought every visible control is of the base type wincontrol.

Visible controls are almost always TControls descendents. TWinControl adds the ability to get focus.
To get capturing back you can for example just doubleclick on a TGroupBox.

Mattias

P.S.: Why does almost every german write my name wrong?

The capturing has still bugs. But you do not need ctrl-c. In most

cases a

doubleclick on a non-wincontrol is enough to get the capturing in

order.

Matthias, what types of controls do you mean with non-wincontrol?? I thought every visible control is of the base type wincontrol.

Jens

Usually you get this when changing the INIT level of a Linux box (possibly a shutdown) and there are no processes left to kill. Normally you can ignore this message.

Michael.

Bug fixed in the editor options dialog. Will soon be available...

The capturing has still bugs. But you do not need ctrl-c. In most cases a doubleclick on a non-wincontrol is enough to get the capturing in order.

Mattias

At first, I don't know what is the state of graphics - TPixmap. Besides, I don't think that pixmaps are prospective way: course, we must be able to read them, but for the compartibility we must be able to read BMP-files. I don't want to use something like Imlib as it will be difficult to make on win32 and other platforms. Besides, using PIXMAP widget can make problems for other interfaces. We should use as least widgets as possible, others can be simply made with platform-independed way. I'll test my TImage implementation on pixmaps and try to load BMP-files. The structure of this file is described (even in Russian), so I'll try.

!!!!!!!!!!!!!!
It is impossible to recompile IDE, so I think, it's good idea to make lcl-oriented ide, using
IDE, included in fpc. It contain debugger, so... I think, its not too difficult....
But now it's not for me.

P.S.: Why does almost every german write my name wrong?

Sorry, this is because I know someone who is written: Matthias.

A lot of question. I think we need a Delphi oracle ... questions left over: 7, 10, 13, 14, 17 at the end of this mail.

There is a new project.pp, so we have finally a valid project format. But the rest of IDE is still using the old prototype code. You can put this to the category "not implemented yet". I think it can be done in a few days.

The component selection, moving and resizing code will be totally replaced. The gtk-interface is still not ready for designing. Hopefully this will be fixed in the next few weeks.

5. The mousepointer changes to a down arrow if I am in the object inspector.

Bug fixed. Will soon be available.

6. The objectinspector is to bit of it is hidden by the maindialog.

I will add object inspector options.

9. A replace-text-dialog is missing for the editor.

I have implemented it yesterday. Will soon be ...

11. It is now impossible to mark lines or words with the mouse, but dbl-click marking of one word is working.

I will check this ...

15. One time I had that in the editor the cursor was shown at a fixed position but I could write at any place.

I got this error too. But to track it down I need a reproducable error.

Mattias

Sorry,

Forget my non-wincontrol rubbish.
TGroupBox is of course a TWinControl descendent, although the workaround is correct. There seems to be no easy way to determine, which component can bring the capuring in order and which not. Thus locating the bug will not be easy. :(

Mattias

P.S.: Why does almost every german write my name wrong?

Probably because the spelling with a 'h' is more common around here. In the Netherlands too.

The interface CRC is calculated after the interface section is read. When things are added/changed in the implementation part. Like here the type names of the arguments.

>>>>>>>>>>>>>>>>> Urspr�ngliche Nachricht <<<<<<<<<<<<<<<<<<

Am 2/20/01, 7:00:42 PM, schrieb "Shane Miller" zum Thema Re: AW: [lazarus] New tests of the IDE:

ability to use keys and the space bar then ours can't either.

However it shouldn't take much to implement. I'm not sure if this is a

priority though.....

It is I think an nice to have feature which could be implemented later, because the searching and selecting in a list is with the keyboard nicer as with the mouse I think.

about it. However, with all the things wrong in the gtkint hopefully when I fix some of them this error will get fixed too.

I get this message since an long time for now.
Which GTK-version and GTK-Pascal-Units do you use??

_________________________________________________________________ To unsubscribe: mail with "unsubscribe" as the Subject archives at http://www.miraclec.com/list_archives/lazarus

Hi

If I compile I get the following message, what is the very long name with the $'s in and why are they differ?

How can a CRC change at this point of compilation where the interfacebase.pp is compiling
the first time?

Compiling interfacebase.pp
interfacebase.inc(22,25) Note: Interface and implementation names are different
_INTERFACEBASE$$_$$_TINTERFACEBASE_$$_CALLWINDOWPROC$TFARPROC$HWND$UINT$LO NGINT$LONGINT =>
_INTERFACEBASE$$_$$_TINTERFACEBASE_$$_CALLWINDOWPROC$TFARPROC$HWND$UINT$WP ARAM$LPARA
interfacebase.inc(136,1) Warning: Function result does not seem to be set interfacebase.inc(248,1) Warning: Function result does not seem to be set interfacebase.inc(314,1) Warning: Function result does not seem to be set interfacebase.inc(345,1) Warning: Function result does not seem to be set interfacebase.inc(381,1) Warning: Function result does not seem to be set interfacebase.inc(422,1) Warning: Function result does not seem to be set interfacebase.pp(121) Warning: ./units/interfacebase.ppu Interface CRC changed -1337124702-301492936
Assembling interfacebase

cu Jens

Answers below:

7. Speedbutton for changing between Form and Unit is not working correctly.

You are correct. I will fix this soon.

10. Can someone remove the executable bits from the lazarus.dci in the CVS?

I will do this too.

13. ListBox: It would be nice if it is possible to scroll with the cursorkeys through the list and select the wanted item with the spacebar.

I'll have to look at the gtk widget.
That's the widget we are currently using so if that doesn't have the ability to use keys and the space bar then ours can't either. However it shouldn't take much to implement. I'm not sure if this is a priority though.....

14. What is with the tooltips, are they disabled again?

Yes, they are disabled again.
I will fix this once I think of a good way of handling them. Someone didn't like the method I was using and I agree with them.

I've never seen this error. I don't get it and therefore can't do much about it. However, with all the things wrong in the gtkint hopefully when I fix some of them this error will get fixed too.

I don't know why you do not get the error. But I know the reason for the bug. The finalizaton sections of synedit are executed before the finalization of forms.pp. This is ok under Delphi and it should also be ok under the lcl. But when a program shuts down Delphi quits/kills all components and after that the program ends and the finalization sections do their work. The lcl does not yet support this clean shutdown and therefore shuts down the components when the forms.pp is finalizing. So, the synedit units are closed before the synedit components are closed. The "bug" exists for a long time. I did not fix it because my priority is on a starting and running IDE, not on a closing one.

Mattias

Has anyone looked at this error, which I posted some days ago ? It is a bit funny ;-)

>>>>>>>>>>>>>>>>> Urspr�ngliche Nachricht <<<<<<<<<<<<<<<<<<

Am 2/19/01, 1:36:42 PM, schrieb Jens Arm zum Thema [lazarus] SEGFault at start of Lazarus:

Hi If you start Lazarus and wait until every dialog at startup is shown

and

change then the console (the splashscreen must be shown, too) (only tested under KDE 2) and when I then change back to the Lazarus

console

after some seconds I get the following SEGFault:

I have tried this but it doesn't crash for me. I start Lazarus and wait for the splash screen to close. Then I go to the console and back to lazarus.

It still works....

Shane

>> 02/28/01 02:42AM >>>

Am 2/28/01, 6:15:16 PM, schrieb "Shane Miller" zum Thema Re: [lazarus] SEGFault at start of Lazarus:

I have tried this but it doesn't crash for me. I start Lazarus and wait

for the splash screen to close. Then I go to the console and back to lazarus.

No you should not wait for the closing of the splash You must start wait until every form is shown and then before the splash close, you must switch to another console and after some time you can switch back and then Lazarus crashes!!!!

It still works.... Shane

>>> 02/28/01 02:42AM >>>

_________________________________________________________________ To unsubscribe: mail with "unsubscribe" as the Subject archives at http://www.miraclec.com/list_archives/lazarus