Mattias Gaertner wrote

One question: How does a 'require=synedit' finds out the gtk requirement? Is there some doc about this feature?

I believe it will take just a little bit more tweeking because it currently doesn't know which to use.

I'll send you what Peter sent me in a private email.

Michael A., will u add them?

We can't at this point. Why? Because it doesn't work.

The idea is to try and use the fpcmake to figure things out via the Package.fpc file. The problem is that fpcmake, at least the only version I can get hold of, looks for the Package.fpc files in the FPC source tree.

That means we would have to make the following directories and files:

\pp\lcl\Package.fpc
\pp\components\codetools\Package.fpc
\pp\components\synedit\Package.fpc
.
.
.

so forth and so on. However we do not want to have Lazarus tied to the my case on Win32 I only use a snapshot of the Win32 compiler with no source and the Lazarus source is locate in a totally different place then the FPC. We don't want to have the two tied together.

So unless the new fpcmake with the '-r' available option can work with the Package.fpc files in the Lazarus source tree we can't use fpcmake and have to stick with putting all of the include directives in the fpcmake.fpc file.

I previously mentioned this to Peter but have not heard back from him yet.

There was also problem with the compiler that wouldn't recompile the interfaces unit for gtk, because it already found the abstract interfaces unit.

I will send you the zip in a private mail

I just heard from Peter and yes the new fpcmake does have a feature to allow this but until he can get me a copy of that we are in limbo.

It wasn't the compiler it was what 'make' was doing with the makefile.

When required i can mail a .zip with updated Makefile.fpc that support this. These files are tested under linux and they produce a correct lazarus executable

Would you email me at least the top level makefile.fpc that as YOU think it should be written. From that I can work with it to find problems that are occuring in Win32 and try to get it and all the rest of the makefile.fpc correct.

At this point I am getting so mixed up with who is changing what and why.

Why is the fpcmake support for lcl removed from the makefiles? With the current Makefile.fpc you need to include a lot of info about the lcl to include. In my Makefile.fpc update this was handled by fpcmake so you could get synedit with only requiring a 'require=synedit' and fpcmake would find out all the other requirements like fcl,regexpr,lcl,gtk etc. for you.

If the problem was that the Makefiles didn't work correct, please try them with an updated compiler dated after 6-jan.

When required i can mail a .zip with updated Makefile.fpc that support this. These files are tested under linux and they produce a correct lazarus executable

Peter

That was the problem. The fpc compilation failed. I was forced to use an old fpcmake to get at least the linux lazarus compilation working.

One question:
How does a 'require=synedit' finds out the gtk requirement? Is there some doc about this feature?

When required i can mail a .zip with updated Makefile.fpc that support this. These files are tested under linux and they produce a correct lazarus executable

Plz send them to me too.

Michael A., will u add them?

Mattias