Lazarus Bugfix Release 4.4

[Martin_fr]

The codetoolmanager errors might be of interest and should be reported with the issue.


IIRC, the "no pascal code found " was mentioned before, but I don't recall details.
Not sure if/how it was resolved.

[geraldholdsworth]

I've opened an issue in the bug reporter with that output, and also a link to the thread I started.

I'm assuming that the "no pascal code found" is due to there not being a project open when I started the application. It's actually the first time I'd seen that, as it doesn't show up when starting from the GUI.

[EganSolo]

I still can’t find a simple reproducible way to report this bug:
Windows 11
Laz 4.4 docked
Two Windows open: the main one with the editor and an assortment of other panels, such as code explorer, project group, project, components, and messages. The second window has only a tabbed editor.

Bug: (minor) While the cursor is in an editor in the second window, the compiler icon (green arrow) is grayed. Further, sorting lines won’t work, meaning the modal dialog for sorting won’t pop up. Workaround: move the file to the primary window.

This bug is not systematic. Some event provokes it, and I haven’t found out just yet what that event is ...

[dsiders]

I still can’t find a simple reproducible way to report this bug:
Windows 11
Laz 4.4 docked
Two Windows open: the main one with the editor and an assortment of other panels, such as code explorer, project group, project, components, and messages. The second window has only a tabbed editor.

Bug: (minor) While the cursor is in an editor in the second window, the compiler icon (green arrow) is grayed. Further, sorting lines won’t work, meaning the modal dialog for sorting won’t pop up. Workaround: move the file to the primary window.

This bug is not systematic. Some event provokes it, and I haven’t found out just yet what that event is ...

Did you miss this part?

Code: Text  [Select][+][-]

1. macOS 15 on an M2 (ARM)

[trom]

Lazarus 4.4 can no longer generate a new LAMW [GUI] Android module (0.8.6.4) Project

Error : List of index (5) out of bound.

I'm on win11, lazarus 4.4 i386 fpc 3.2.2

Best regard

[jamie]

You need to open the sources and track it.

It's possible things got recompiled with RangeCheck on and there is code that goes outside a range check but would otherwise be ok.

 Or there is some code that didn't predict a shortens of an array due to some resources unexpectedly not available.!

Jamie

[Martin_fr]

I still can’t find a simple reproducible way to report this bug:
Windows 11
Laz 4.4 docked
Two Windows open: the main one with the editor and an assortment of other panels, such as code explorer, project group, project, components, and messages. The second window has only a tabbed editor.

Bug: (minor) While the cursor is in an editor in the second window, the compiler icon (green arrow) is grayed. Further, sorting lines won’t work, meaning the modal dialog for sorting won’t pop up.

The icon, as well as the enabled state of "sort selection" (only enabled if a selection exists) are updated in OnIdle handlers  (and, not sure, maybe only in the past, when the main-ide-bar got focused??).
Calling "sort" (even via keyboard) might do nothing, if the item is disabled.

So one possibility would be that the idle handler is blocked. But no idea why/how => unless your system is running at constant load (not 100% but one cpu core at 100, so depending on how many cores you have, that is a different percentage).

The updates also (maybe??) require that there was user input. There is a handler too, and it sets a flag. Every keypress and mouse move should trigger that... Again: should. Again: no idea.

[EganSolo]

The icon, as well as the enabled state of "sort selection" (only enabled if a selection exists) are updated in OnIdle handlers  (and, not sure, maybe only in the past, when the main-ide-bar got focused??).
Calling "sort" (even via keyboard) might do nothing, if the item is disabled.

So one possibility would be that the idle handler is blocked. But no idea why/how => unless your system is running at constant load (not 100% but one cpu core at 100, so depending on how many cores you have, that is a different percentage).

The updates also (maybe??) require that there was user input. There is a handler too, and it sets a flag. Every keypress and mouse move should trigger that... Again: should. Again: no idea.

Martin: Perhaps this bug might shed more light?
When I’m in the second editor and hit CTRL+R to perform a find-and-replace, the IDE switches back to the primary editor. It’s as if some logic isn’t checking for the existence of a second editor window in anchored mode. It seems to me that this bug is not related to the On-Idle mechanism ...

[Martin_fr]

Martin: Perhaps this bug might shed more light?
When I’m in the second editor and hit CTRL+R to perform a find-and-replace, the IDE switches back to the primary editor. It’s as if some logic isn’t checking for the existence of a second editor window in anchored mode. It seems to me that this bug is not related to the On-Idle mechanism ...

If the IDE for some reason acts on the primary editor, then if that has no selection then "sort selection" would not be enabled.
I would guess that if you have a selection in both, and if you are in the 2nd, and if you invoke "sort selection" then it sorts the first.

There was an issue that certain actions in the 2nd would switch the tab in the 1st editor window. That was caused by the embedded/docked form editor. https://gitlab.com/freepascal.org/lazarus/lazarus/-/issues/39417
That was merged to 4.4

It starts getting hard to track the issue in this mega thread. Did you test with/without docked form editor?
Was this issue present in 4.2 ?

Ultimately it needs to be reported.
I can add any ideas I have on this thread. But that won't fix it, and I don't currently work on the docking stuff.

[BSaidus]

Hello.
I have a situation since the 4.0 release of lazarus (exists on all 4.x releases).
I allways do recompile lazbuid, startlazarus, lazarus in release mode (optimized), and when starting the IDE using startlazarus.exe, the splash screen do not appear.
PS: in

Code: Pascal  [Select][+][-]

1. lazarus.cfg there is only the --pcp='configpath'
2.  

So, What happen with startlazarus splash screen ?

[Martin_fr]

Just checked (don't know your OS, but mine Windows 64bit), and it seems startlazarus is indeed not showing the splash.

Haven't checked why... I recommend to open a bug report.

[BSaidus]

Just checked (don't know your OS, but mine Windows 64bit), and it seems startlazarus is indeed not showing the splash.

Haven't checked why... I recommend to open a bug report.

Me too, Windows 11 x64, but the IDE & others are 32bit.
I do not have an account in gitlab, so if someone kindly can do that.

[calebs]

Hello all. Many thanks for the good work with lazarus i am very grateful to all the comunity members of this.
I've found long time ago and always forget to mention it don't know if somebody has noted this  (bug i think)
Drop a page control to a form, then proceed to create many pages with captions that exceed the available form space for displaying all tabs and then proceed to use the arrows buttons from above right to scroll betwen pages and the right button advance to the next tab (right) and if i keep pressing it it keeps moving to the right. But when the first tab gets out of sight or when i get to the last tab (page) the arrow buttons are disabled and doesn't work anymore.
if i change the active page value from the object inspector to the first page again those buttons are enabled again but it repeats the same behaviour when i get to the last again.
I think this is happening from several versions of lazarus before but forgot to mention it.
Thanks!

[BSaidus]

Just checked (don't know your OS, but mine Windows 64bit), and it seems startlazarus is indeed not showing the splash.

Haven't checked why... I recommend to open a bug report.

bug created.
https://gitlab.com/freepascal.org/lazarus/lazarus/-/issues?show=eyJpaWQiOiI0MTk0NyIsImZ1bGxfcGF0aCI6ImZyZWVwYXNjYWwub3JnL2xhemFydXMvbGF6YXJ1cyIsImlkIjoxNzgzNjM1MTR9

[BrunoK]

[Edit] : wrong discussion. Lots of disruption in forum access today.
We are losing our time, OP showed code that was in trunk before 2 months ago, current trunk has different code.