Hi,

For years I have been using the same script to compile FPC, I have multiple version of FPC installed in the same machine, mainly my script do:

make -j 9 clean all

make -j 9 all install FPMAKEOPT="-T 9"
INSTALL_PREFIX=.../fpclazarus/fpc/cmp/3.3.1 OPT="-O- -O1 -gl" PP=.../fpclazarus/fpc/lib/3.2.2/ppcx64

I have been a few months without compiling FPC trunk, but when I tried it today, I got the next message after calling "make -j 9 clean all"

make: -iVSPTPSOTO: Command not found Makefile:235: *** The Makefile doesn't support target -, please run fpcmake first.  Stop.

I just run fpcmake before call "make clean all" and "make -j 9 all install", and everything works as expected.

My question is, until now I have never needed to call fpcmake before compiling trunk, it is a requirement now to execute fpcmake or there is something wrong in my side?

My system is Xubuntu Linux x86_64.

Thanks.

Best Regards.

Victor Campillo wrote

> make: -iVSPTPSOTO: Command not found > Makefile:235: *** The Makefile doesn't support target -, please run > fpcmake first.  Stop.

Binary "fpc" not found. Maybe a different account or something else changed in your $PATH ?

Victor Campillo wrote

For years I have been using the same script to compile FPC, I have multiple version of FPC installed in the same machine, mainly my script do: make -j 9 clean all

This is wrong. You cannot safely run "clean" and "all" in parallel. You have to run them separately.

Jonas

This indicates that the PP variable is not valid ?

I run make clean all almost daily, I have not seen what you report here.

Michael.

I forgot to reply to this: as Michael mentioned, this does not mean fpcmake needs to be run. Even after running fpcmake, the target "-" would still not be supported.

The issue is probably that the ppc or fpc binary in the compiler subdirectory was in your path, and then got deleted by "make clean" while in parallel "make all" tried to use it for compilation.

Jonas

Hi,

I can reproduce this, or something very similar for the clean/distclean targets, if I have no FPC installed on my machine, but I specify PP= as a Makefile argument to a "bootstrap" compiler binary, and I use OS_TARGET= as well, for crosscompiling.

For example, if I try this on Linux/x86_64:
make clean OS_TARGET=aros PP=/path/to/bootstrap/ppcx86-3.2.2

I get a lot of messages like:

ake -C compiler distclean
make[1]: Entering directory '/home/charlie/Develop/fpc/source/compiler' make[1]: -iTP: Command not found
make[1]: -iTP: Command not found
make[1]: -iTP: Command not found
make[1]: -iTP: Command not found
make[1]: -iTP: Command not found
make[1]: -iTP: Command not found
/bin/rm -f ppcrossx64 ppc ppc1 ppc2 ppc3 ./msg2inc pp1.wpo pp2.wpo

So there's definitely something fishy with the bootstrap/PP compiler use, if the build box has no FPC installed otherwise. Which to my understanding, it shouldn't be needed, if PP is specified?

Interestingly, if I don't specify OS_TARGET, the problem doesn't occur.

Note that for a real build ("all" or "crossall" target), the specified PP compiler still gets used as expected.

Charlie

Hi,

Thank you very much to everyone for the quick answers.

If I specify the option PP=.../fpclazarus/fpc/lib/3.2.2/ppcx64 to the command "make -j 9 clean all", the it works as expected. Probably until now, it worked just by pure lucky.

But just to make sure my script to build FPC is safer, is this is ok?

make -j 9 clean PP=.../fpclazarus/fpc/lib/3.2.2/ppcx64

make -j 9 all install FPMAKEOPT="-T 9"
INSTALL_PREFIX=.../fpclazarus/fpc/cmp/3.3.1 OPT="-O- -O1 -gl" PP=.../fpclazarus/fpc/lib/3.2.2/ppcx64

It works, but I want to make sure that this is the right way to build FPC.

P.D. :Thank you very much to the FPC team for this great project.

Best Regards.

Victor Campillo

[...] OPT="-O- -O1 -gl"

Is this paranoid or useful?
-O- -O1

Mattias

Try the same, but use FPC= instead of PP= . Let it point to the ppc binary.

Hi,

It makes no difference.

Charlie

Mattias Gaertner wrote

Is this paranoid or useful? -O- -O1

Good point. My script have many years and I don't remember why I used this options or if I gathered them from someone else, but reviewing it now, it does not seems logical. I suppose -O1 overwrite -O-, so specify -O- is completely useless.

Hi,

No, it isn't. -O<number> parameters will _enable_ certain optimizations. This means, -O3 -O1 will not switch the compiler "back" to -O1, just stay at -O3, because at -O3 level, all -O1 optimizations are already enabled, there is nothing to enable.

-O- however will disable all optimizations. So if you want to disable all previously enabled optimizations, you indeed need to go to the -O- -O1 route. So it is actually useful.

(The above is quite easy to verify, see options.pas, line 2350 and below in the compiler sources.)

Charlie

Good to know.

Now I am pretty sure that I copy that options from someone else, because with my knowledge I could not figure out that myself.

Thank you Charlie.

Best Regards.