The good news is that it builds with FPC 3.0.0 on:

PPC Linux, Debian "Lenny"
Raspbian "Jessie" on RPi2
SPARC OpenSXCE 2014 (approx. Solaris 11)
SPARC Linux, Debian "Squeeze"
x86 linux, Debian "Lenny" and "Jessie"
x86_64 linux, Debian "Jessie"

I've not tested any other ARM variant since I can only run those on minimal hardware or using Qemu which is infeasibly slow. I've also not tested MIPS where Lazarus is known to be a problem.

An annoyance affecting all platforms above is that the timeout in the "Configure Lazarus IDE" -> "Compiler" box is impossibly fast, if the default isn't right about the only way to use it is to paste a prepared name.

Slightly more serious from my POV is that a project that created a form early to set up a configuration file from a template is now giving problems on all tested platforms. I'm assuming that http://free-pascal-lazarus.989080.n3.nabble.com/Lazarus-A-change-in-LCL-require-your-TCustomForm-descendants-to-have-a-resource-or-to-use-a-CreateNer-td2402646.html is relevant, but at present am struggling.

Impressive. :)

What timeout?

This is a pretty old change.
Simply use CreateNew instead of Create.

Mattias

Thanks, I try to exercise things on a fairly regular basis. Help probably won't work on SPARC and PPC Linux due to issue 22696 (FPC), and there might be one unit excluded from Solaris which would be useful... I'll check this.

After the compiler name has been changed by the addition or removal of a single character, it goes and checks whether the filename is now valid. The timeout here is something like 10th sec, and displays a dialogue if the name is bad. The FPC sources etc. fields are OK, but this one will cause problems for anybody trying to use multiple compilers by specifying e.g. ppcppc-3.0.0 rather than just fpc (there's nowhere to put in a -V option which could be picked up by fpc).

I noticed that one a few months ago but assumed somebody else was aware: I know, there's no need to tell me.

Yes, but has only now become mandatory.

Simply use CreateNew instead of Create.

I haven't had any success with that yet, but started from a position of suspecting that the problems I was seeing were platform-specific, it was quite a while before I found that some systems worked because they already had the configuration file that was being created.

Fixed.

The change was already in 1.0.

> Simply use CreateNew instead of Create.

Or set RequireDerivedFormResource to false.

Mattias

Yippee :-) Hopefully that's something that can get into the release.

The unit/form is outside the project but is predefined (i.e. not constructed at runtime). It's created early (before the main form's constructor has completed) in order to check and if necessary create a configuration file. Please correct me, but since its layout is stored in resources I think CreateNew() doesn't fix anything.

I'm still not entirely convinced, but I think the problem was that it had a {$I unitname.lrs} in the initialisation part rather than a {$R *.lfm} at the start of the implementation.

So I'm fairly sure this isn't an IDE etc. problem, but since I can't easily say what version created these files it might be one that will bite more people with elderly projects.

[...] Yippee :-) Hopefully that's something that can get into the release.

Yes.

The link you gave is about forms without resources. If your buggy form has a resource, then this was misleading.

This is not enough detail to help you.

Mattias

Minor widget set issues.

On a recent Raspbian running on an RPi2, using a directly-connected HDMI screen with USB keyboard and mouse, with OpenBox or LXDE as window manager:

GTK2 radio box squashed

Qt messagebox button icons garbled.

The former case suggests that some form sizing etc. might not be working quite as expected. It's quite possibly a layout error on my part, but so far it only shows up on Raspbian+OpenBox although I've not yet checked OpenBox on other platforms.

In the second case, the icons on a "do you really want to quit" type messagebox both contain the same noise pattern. Again, I only see this on Raspbian but in this case it affects both OpenBox and LXDE.

In both cases, this has to be a directly-connected screen: executing over SSH etc. looks fine.

No obvious problems on Debian Jessie running on x86_64, i.e. they're probably Raspbian-specific.

Also on straight Raspberry Pi Debian (i.e. not Raspbian) from http://sjoerd.luon.net/posts/2015/02/debian-jessie-on-rpi2/

An earlier attempt at this wouldn't compile Lazarus, this time I used the notes at https://www.debian.org/releases/jessie/amd64/apds03.html.en as a checklist for what needs to be on the system, I think the really important thing was the console-setup package which looks like it's needed even if the idea is to run the system headless.

I am not sure if you saw this :
http://forum.lazarus.freepascal.org/index.php/topic,30500.msg194271.html Apparently FPC works with MIPS but Lazarus does not. Do you have experience with MIPS?
I may join the effort later and get a MIPS Creator to see why Lazarus does not work.
However I am quite a newbie in compiling FPC for different architechtures. Let's see...

Anyway, I appreciate your work with unusual platforms and CPU endiannes issues + patches.
Different gadgets and so called TV-boxes can do personal computing more and more in future. Their CPU and OS selection varies a lot. BTW, I personally have an Intel-based passively cooled ASRock Beebox as my main computer now. I am amazed every day about this gadget!

Regards,
Juha

I've mentioned this problem in the past unfortunately, but I'm slightly hobbled here in that while I've got a couple of SGI systems with IRIX the only MIPS Linux I'm able to run is based on Qemu. Two or three years ago one of the vendors here promised to send me a netbook but it didn't materialise, and quite frankly I have very limited resources when it comes to going out and buying new hardware just in case it turns out to be useful.

I may join the effort later and get a MIPS Creator to see why Lazarus does not work. However I am quite a newbie in compiling FPC for different architechtures. Let's see...

I think a Creator is definitely worth looking at very carefully, and I'd add that MIPS is aware of FPC even if not currently (as far as I know) on the list or active in the forum. The RPi has some nice features but once one starts to do real work with it some decidedly dodgy ones become apparent, like almost all I/O traffic going through a single USB connection.

As a general point, for anything like this it's worth having an FTDI (or whatever) USB -> 5V serial adapter so that if it has a pre-boot menu etc. you can get to it.

Anyway, I appreciate your work with unusual platforms and CPU endiannes issues + patches.

Thanks :-) The way I look at it is that some architectures- e.g. Linux SPARC- exercise the code generator a bit more than others on account of things like alignment requirements- hence in particular Mantis 22696 that I mentioned a few days ago. Looks like SPARC-64 Debian will be usable imminently...

Different gadgets and so called TV-boxes can do personal computing more and more in future. Their CPU and OS selection varies a lot. BTW, I personally have an Intel-based passively cooled ASRock Beebox as my main computer now. I am amazed every day about this gadget!

Agreed. Meanwhile I'm still tinkering with my mainframe emulator: 1MHz CPU clock, 4-6 uSec core memory and so on. It really does emphasise how far we've come :-)

Waiting for various things to happen here, but meanwhile http://xkcd.com/679/ enjoy.

Thought not 1.6 specific ther'is an annoying bug with the IDE undo. If you
create an  event for instance
procedure TForm1.Button1Click(Sender: TObject);
begin
end;

do ctrl+z

and only Button1Click from source is removed and not its reference in the OI. So if you try to recreate the Button1Click event via OI or dbclick on button1 the IDE is give a error msg that didnt help to find the problem (unit section enlighted)

Salvatore

Ok "although" instead of "Thought"!
Ok wrong tread!
Should be opened an issou in the BT?
thanks for the huge amount of replies!

2015-12-26 10:00 GMT+01:00 Salvatore Coppola :

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151227/f479b007/attachment.html>

Salvatore Coppola wrote

Ok "although" instead of "Thought"! Ok wrong tread! Should be opened an issou in the BT? thanks for the huge amount of replies!

I don't know if it is even possible and wanted to extend the source editor undo function with the ability to change the assigned designer. So yes you can open an issue report in the BT if there isn't the same already but don't expect to get a solution soon. I even think this behavior will be identified "as intended" and your issue as "won't be resolved". Unless you supply a good patch, of course :)

Ondrej
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151227/36103732/attachment.html>

Salvatore Coppola wrote

the IDE is give a error msg that didnt help to find the problem (unit section enlighted)

I find the error message clear enough:

and in the messages window you find

unit1.pas(1,11) Error: identifier not found: Button1Click

So there is no do you want to change?
- do you want to remove the "Button1Click" reference in OI on undo? or
- do you want to create Button1Click even if there is reference in OI but the method cannot be found?

Ondrej
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151227/21b7c6a5/attachment-0001.html> -------------- next part --------------
A non-text attachment was scrubbed...
Name: hhfadajh.png
Type: image/png
Size: 4308 bytes
Desc: not available
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151227/21b7c6a5/attachment-0001.png>

2015-12-27 22:29 GMT+01:00 Ondrej Pokorny :

Yes, this can be a good option because in the other case (compiling) the IDE suggest to self remove "Button1Click" in the OI. Salvatore

Ondrej

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151228/0243953f/attachment.html> -------------- next part --------------
A non-text attachment was scrubbed...
Name: hhfadajh.png
Type: image/png
Size: 4308 bytes
Desc: not available
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151228/0243953f/attachment.png>

Which one?

Ondrej
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151228/f64c899f/attachment.html>

Create the
procedure TForm1.Button1Click(Sender: TObject);
begin
end;
Instead of the error message

2015-12-28 9:16 GMT+01:00 Ondrej Pokorny :

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151228/b133ea0c/attachment.html>

Please fill an issue report in the BT.

Ondrej
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151228/eb28b37f/attachment-0001.html>

Done

2015-12-28 9:50 GMT+01:00 Ondrej Pokorny :

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20151228/908440e6/attachment.html>