I'm trying to compile Lazarus on FreeBSD using 1.0.5 (of today), and am having troubles:

- osdirseparator doesn't exist, but its purpose can be easily guessed.
- TProcess is used in two ways, while it has only one constructor, and TProcess is only defined in the fcl process function?

   One is   <var>:=TProcess.Create(NIL) (Arg probably TComponent according
   to the fcl?) and the other is TProcess.Create(<filename>,[flag1,flag2]);

What is going on here? I first thought Lazarus was simply out of sync with 1.0.5 as it is now, but the sources seem to use two different definitions of Create()?

If I simply go to lazarus/ and type "gmake", I get errors on old style (with string, options) styled params.

Since this afternoon, another problem was "added":

gtkint.pp(125,1)
There is no method in an ancestor class to be overriden: TGTKOBJECT.DESTROYCARET(LONGINT)

Strange. This is used in lazarus for a long time and it is defined in sysutils for every OS.

The TProcess.Create(nil) belongs to the new TProcess which is a TComponent. Lazarus will switch to the new TProcess, as soon as we got a new fpc rpm. Marc's TGDBDebugger is already using the new TProcess, and therefore can't be compiled. AFAIK Marc is working on the gdb support.

Mattias

Seems to be replaced by "Pathdelim"

Is there any way to build Lazarus now? E.g. without deugger?

Marco van de Voort wrote

gtkint.pp(125,1) There is no method in an ancestor class to be overriden: TGTKOBJECT.DESTROYCARET(LONGINT)

You need to do some form of 'make clean'. I had this same error this afternoon when I did and update but not a 'make clean'. As soon as I did that then everything compiled again.

Forget this one. I probably did an update, and didn't clean afterwards.

The odd uses of TProcess are in definetemplates.pas. e.g. line 1098.

You're right. I have problems accessing the fpc cvs (permission denied). Hopefully they will fix it soon.

Is there any way to build Lazarus now? E.g. without deugger?

The TGDBDebugger is currently not used. All TProcess use the old style. You need the old OSDirSeparator to compile. We need a newer fpc rpm.

Mattias

Hi,

I have noticed there is a lot of progress on Project Lazarus. This is great news!!

I have downloaded lazarus_ide_installer.exe but when I install and run Lazarus I don't see any change.
I am assuming this installer has not been updated - still Lazarus 0.7alpha. Why is this, resource issue, bugs..?

Kind regards,
Ronan van Riet

Hi, I have noticed there is a lot of progress on Project Lazarus. This is

great

0.7alpha.

Why is this, resource issue, bugs..?

The last I heard, the nightly builds for Windows haven't been updated since August because of the compiler requirements. Obviously this is still the case since the version is 0.7 instead of 0.8.1.

I guess if you really want to experiment with the new and fixed stuff, you can compile it yourself.

I was just trying to get the nightly builds going again this weekend but hit another snag. There is an imcompatibility with GTK-Win32. We are using a construct that is available in Linux but not in Win32. It is being looked at however it might mean that we have to really develop the Win32 API. However at this point that also had some build problems.

I've been working on the Win32-API interface for the last week or so and have made a little progress (it compiles again, common dialogs work, I've fixed minor problems with edit, memo, radio buttons, and check boxes, ...).

I still haven't got events working. I've done some experimentation and the only thing that stands between it working is associating only certain messages with an object. I tried using a TList but I was getting access violations for no apparent reason. I guess the only way is to use dynamic arrays whenever 1.1 becomes usable with Lazarus unless someone else has a brilliant idea.

The IDE still isn't working, but it gets further before failing. And on a similar note, I have no idea the of the Win32 equivalents of the arrow and calandar components. Right now, I'm using cursors for arrows and for calendar, I'm opening the time/date control-panel applet and using calls to SetLocalTime and GetLocalTime. I know these are wrong but they were the closest approximations I could think of.

I'll probably have a new update that you guys can add to the official source in the near future. At that time, I guess you and whoever else wants to help can help fix the bugs, add the new features, and re-implement the stupid ideas I've had (like the arrow and calendar components) in a better way.

And on the subject of GTK+ on Windows: I'm not particularly fond of it but I think it would be a shame if it were no longer workable. It would just be another nail in the coffin of being able to use the API of your choice.

I'm sorry it's taken me so long to reply but I've been kind of busy lately.