Code Prettifier / Formatter

[Fibonacci]

There is JEDI, but its old, not updated, and barely works.

The most suitable for me I found is this one, but its not enough:
https://gitlab.com/rchastain/pcf

So are there any recommended code formatters to keep some standard while working in a team?

[DomingoGP]

Have you tried the version of Jedi code format that comes with Lazarus?

components\jcf2\CommandLine\Lazarus\jcf.lpr

It is up to date and I think it works quite well (although my opinion is somewhat biased).

[Thaddy]

And it is also maintained and best suited for Lazarus.

[Fred vS]

There is also xtop, it is a fork of fpc ptop with extended feature and without memory-leak.

[Fibonacci]

Good, thanks, Im collecting answers and will test them all later Thanks and I hope for more alternatives.

[Bogen85]

There is JEDI, but its old, not updated, and barely works.

The most suitable for me I found is this one, but its not enough:
https://gitlab.com/rchastain/pcf

So are there any recommended code formatters to keep some standard while working in a team?

The command line JCF (Yes.... JEDI...) does get updates, but it breaks a lot...

Recently the updating to for lowercasing on a lot of things broke, and I need to find out what git commit broke things, and file an issue...

See https://gitlab.com/freepascal.org/lazarus/lazarus/-/issues/41147 on how to build it from the command line.

You have to (well, you don't.... it is easier to the way) go into Lazarus to setup the initial xml config file, but after that you can edit it by hand...

I'll take a look at PCF.

[Bogen85]

Have you tried the version of Jedi code format that comes with Lazarus?

components\jcf2\CommandLine\Lazarus\jcf.lpr

It is up to date and I think it works quite well (although my opinion is somewhat biased).

Yes! And you do fix issues with it! Thanks!

It does work well. The only current problem I have (and I need to file an issue) is that forcing of specific types to lower case stopped working, and I'm not sure when.
I looked at that section of code and it had been refactored, but I'm not sure if that is when it broke.

I will file a report.

[Bogen85]

The command line JCF (Yes.... JEDI...) does get updates, but it breaks a lot...

"breaks a lot" is not completely accurate.
It could maybe use a better pipeline job.

But I should not complain, for the most part it works quite well for me, and the only one for me that works satisfactorily.

[Roland57]

The most suitable for me I found is this one, but its not enough:
https://gitlab.com/rchastain/pcf

If you suggest some improvements, maybe I could take a look. (I am the maintainer of the project.)

What does it lack?

[Fibonacci]

I would need some time to check all the little things and make a report, but Ill give you an example right away:

Original code:

Code: Pascal  [Select][+][-]

1. unit Unit1;
2.  
3. {$mode ObjFPC}{$H+}
4.  
5. interface
6.  
7. const
8.   CP_ACP     = 0;     // default to ANSI code page
9.   CP_OEMCP   = 1;     // default to OEM (console) code page
10.   CP_UTF16             = 1200;  // utf-16
11.   CP_UTF16BE           = 1201;  // unicodeFFFE
12.   CP_UTF7 = 65000; // utf-7
13.   CP_UTF8 = 65001; // utf-8
14.   CP_ASCII   = 20127; // us-ascii
15.   CP_NONE    = $FFFF; // rawbytestring encoding
16.  
17. procedure fpc_ansistr_assign2(   var dests:pointer;s2:pointer); stdcall; compilerproc;
18.  
19. implementation
20.  
21. procedure fpc_ansistr_assign2(var dests: pointer; s2: pointer);    stdcall; [public    , alias: 'FPC_ANSISTR_ASSIGN2'];
22. begin
23. end;
24.  
25. procedure fpc_ansistr_assign2(var dests: pointer; s2: pointer); [  external name   'FPC_ANSISTR_ASSIGN2'];
26.  
27. procedure test_empty;
28. begin
29. end;
30.  
31. procedure test_short;
32. var
33.   a: shortstring;
34. begin
35.   a := '----';
36.   writeln('short = ', a);
37. end;
38.  
39. end.

PCF formatted:

Code: Pascal  [Select][+][-]

1. unit Unit1;
2.  
3. {$MODE ObjFPC}{$H+}
4.  
5. interface
6.  
7. const
8.   CP_ACP = 0; // default to ANSI code page
9.   CP_OEMCP = 1; // default to OEM (console) code page
10.   CP_UTF16 = 1200; // utf-16
11.   CP_UTF16BE = 1201; // unicodeFFFE
12.   CP_UTF7 = 65000; // utf-7
13.   CP_UTF8 = 65001; // utf-8
14.   CP_ASCII = 20127; // us-ascii
15.   CP_NONE = $FFFF; // rawbytestring encoding
16.  
17. procedure fpc_ansistr_assign2(var dests: pointer; s2: pointer); stdcall; compilerproc;
18.  
19. implementation
20.  
21. procedure fpc_ansistr_assign2(var dests: pointer; s2: pointer); stdcall; [public, alias: 'FPC_ANSISTR_ASSIGN2'];
22. begin
23. end;
24.  
25. procedure fpc_ansistr_assign2(var dests: pointer; s2: pointer); [external name 'FPC_ANSISTR_ASSIGN2'];
26.  
27.   procedure test_empty;
28.   begin
29.   end;
30.  
31.   procedure test_short;
32.   var
33.     a: shortstring;
34.   begin
35.     a := '----';
36.     writeln('short = ', a);
37.   end;
38.  
39. end.

So the first thing is the align in the consts (JCF aligns that fine). The second thing is the weird indent after "external" procedure.

I would make a much longer list with little things but it would take me some (long) time. Are you willing to fix it all?

[Thaddy]

It can be done, but it looks like it needs a huge look ahead buffer, because you do not know beforehand where to align.
(Or back prop, but that is the same, only slower, worst case you need both, on block level)

[Roland57]

So the first thing is the align in the consts (JCF aligns that fine).

Thank you for the example.

The second thing is the weird indent after "external" procedure.

I never saw function headers like that. 

I would make a much longer list with little things but it would take me some (long) time. Are you willing to fix it all?

I cannot promise to fix all, because I am not sure of success, but I promise to take a look.

By the way, there is a bug that I have been hunting for a long time and that I would glad to fix.

Both versions of the GUI (LCL, MSEgui) would also need to be finished.