Announcment: RC2 for the upcoming Version 1.2.

[snorkel]

I got it working, I had to remove the laz report component from the datamodule and recreate it.
Works fine now and I can't see any difference between the one I deleted and the the new one I dropped on the form.
For whatever reason it could not stream the saved report to the form for storage.
oh well works now.

[snorkel]

I just noticed that the build number auto increment works like it should in 1.2RC2 :-)
In Laz 1.x the number that got stored to the resource was 1 less than what the project properties said.

Looking forward to the final release :-)

[Martin_fr]

I am fine with it just being in the title caption as well.  Just ideas :-)

Lazarus 1.3 Revision 43949

Must be enabled in options first. (Because other such options are off by default too, like showing the path). Default is a "cleaner" title.

[JuhaManninen]

I just noticed that the build number auto increment works like it should in 1.2RC2 :-)

Can you please look at this:
  http://bugs.freepascal.org/view.php?id=25683
I will look at it, too, but now I have something else going on...

[Windsurfer]

I am using an i7 PC with Linux Mint 16, Cinnamon. [EDIT -- 1.2RC2]

Jedi Code Formatter halts at conditional defines like those below.

Code: [Select]

{SIfdef Win32 or Win64}
  Do something...
{$Endif}
{SIfdef Linux}
  Do something else...
{$Endif}

The conditionals are defined in Project | Project Options | Other

If I comment out the {$Ifdef } statements then the JCF works.
Am I doing something wrong?

[snorkel]

I just noticed that the build number auto increment works like it should in 1.2RC2 :-)

Can you please look at this:
  http://bugs.freepascal.org/view.php?id=25683
I will look at it, too, but now I have something else going on...

The issue I had in Laz 1.x was the version number stored in the application resource was always one less than what the project options said when the auto inc option was enabled, it was fine when you manually incremented it.  Not sure if that bug report is the same issue.

Seems to work fine in 1.2rc2 though.

[JuhaManninen]

Seems to work fine in 1.2rc2 though.

Only when using Shift-F9 (Build). Updated the issue.

[Windsurfer]

I think I have finally isolated a problem with the Jedi Code Formatter. 
(1.2RC2, Intel i7, Linux Mint 16, Cinnamon, 64 bit, GTK2)

I submitted a rather garbled bug a few days ago, and concluded I had made an error. Could someone please test and confirm the code below, and I will resubmit it.

When an If ... Then block is split with an {$Ifdef}...{$Endif} construct, the compiler handles it, but the JCF does not. The following code snippets illustrate this.

Create a project with a single form.

In Project/Project Options/Other add the conditionals 'Red' and 'Blue'. (Case should not matter.) Select one of the conditions. leave the other unselected.

This works in JCF:

Code: [Select]

procedure TForm1.FormClick(Sender: TObject);
var
  myBool: boolean;
begin
  MyBool := True;
 
  {$IfDef Red}
   If MyBool then Form1.Caption := 'Red'
   else Form1.Caption := 'Black';
  {$Endif}
  {$IfDef Blue}
    If MyBool then Form1.Caption := 'Blue'
   else Form1.Caption := 'Black';
  {$Endif}
end;

This does not work in JCF:

Code: [Select]

procedure TForm1.FormClick(Sender: TObject);
var
  myBool: boolean;
begin
  MyBool := True;

  if  MyBool
  {$IfDef Red}
     then Form1.Caption := 'Red';
  {$Endif}

  {$IfDef Blue}
     then Form1.Caption := 'Blue';
  {$Endif}

end;

[Edit] Adding  an Else after the blocks also appears not to work very well.

[snorkel]

Seems to work fine in 1.2rc2 though.

Only when using Shift-F9 (Build). Updated the issue.

Ok, it does seem to only increment it in the exe if you do a build and then a compile right after. 
Just doing the build increments it in project options, but not in the exe resource.

Prior to 1.2 I don't remember it incrementing in the exe if I did a compile right after the build, it would never show the correct version number in the exe in 1.x from what I remember.  I had to turn off the auto inc feature and do it manually.

[snorkel]

More problems with Lazreport.
I had a issue with lazreport on a project I opened up with 1.2RC2 before with a stream read error, and I just got it again on
a more substantial older project.

This project has a couple of reports and one is on a form and it worked just fine, the other one is on a datamodule and it gives the
stream read error when I try and save the project.  The report on the datamodule lost all the objects in the bands when I opened it as well.

The other project I had the problem with also had the report on a datamodule, so the problem with the stream error seems to be if the Lazreport from 1.0.14 etc was on a datamodule.


UPDATE:

I restored a older copy of the datamodule with the report on it from my SVN repository and the lfm file from the 1.2rc2 has totally messed up the reportxml that lazreport uses and appears to have converted it to a binary form.  The lfm from before I opened it with 1.2RC2 had a report.xml with text based XML in the LFM file.   It appears that when a 1.x project is opened with 1.2rc2 it's doing something to the data module LFM file and pretty much wiping out the stored LazReport XML.

I can't be the only one that has used Lazreport by putting the report on a datamodule.  I will see if I have any other projects from 1.x that I put the report on a datamodule. 


To get the XML back I had to open a older copy of the project on 1.x then copied the report to a form in the project, I then copied that form to the 1.2 copy of the project and opend it, then copied the good report from the form to the datamodule.

UPDATE:

The report on the datamodule was stored as reportform and was binary.
The report on the form was stored as reportXML and was in the LFM as plain text XML.

Not sure but it might be the version of LazReport shipping with 1.2RC2 does not know what to do with reportform and binary data in the LFM.
It seems that in 1.x Lazreport was storing the report on a form differently than on a datamodule. 

Has anyone else noticed this behavior?

Update 3:

seems to be related to this:
http://bugs.freepascal.org/view.php?id=24756#bugnotes

In my case it seems the reports I had on forms where converted in 1.x but not those on datamodules.

[tintinux]

Hi

I have some remarks and question, only about the cross compiler installation.

1)
The install title, displayed several time during install is "Lazarus" which might be confusing. 'Lazarus cross compiler" or something more specific would be better and might prevent mistakes.

2)
I installed Lazarus in "c:\Lazarus\1.2RC2" (not the default location, but could be). When choosing the installation location of the cross compiler, I browse to the same directory, but the path displayed after choosing it is "c:\Lazarus\1.2RC2\lazarus". Is it a correct choice ? If we confirm the default location and do not browse for another, the installation folder is set to the same directory "c:\lazarus".

3)
In addition, and if possible, it would be smarter if the cross compiler was for ISS "a patch" of the corresponding version of the IDE.  I do not know if ISS have such functionnality, but with other product I uses this allows to remove the question for an install location and reduces risks of mistakes
.
Best regards
Tintinux

[Martin_fr]

1 and 2: please report on mantis.

2: You need to remove the extra "lazarus"
IIRC the main installer, adds the same, if you change location.

3: If it did not ask then you could not install it with the secondary install.
Keep in mind, that you can have several installations of the same version.

Yes there could be checks, if the selected folder contains an installation. Patches for this will be welcome.

[valdir.marcos]

Will there be a Lazarus 1.2 RC3?

http://wiki.lazarus.freepascal.org/Lazarus_1.2_fixes_branch#Roadmap_to_1.2

Is it possible that Lazarus 1.2.0 FINAL to be released around 2014-03-01?

http://bugs.freepascal.org/roadmap_page.php says:

Lazarus - 1.2.0 (Scheduled For Release 2014-03-01) = 98%
Packages - 1.2.0 (Scheduled For Release 2014-03-01) = 100%
Patches - 1.2.0 (Scheduled For Release 2014-03-01)  = 100%

[Martin_fr]

This is actually an open decision, that will probably be made in the next 7 days.

*IF* there is an RC3, then release will be a month later.

[snorkel]

take the time needed to do a great release.  I would rather see a RC3 and wait a bit for a 1.2 final.