Hi list,

I wonder if the problems below are caused by wrong make files or problems in fpcup:
(fpc trunk, Windows; fpcup compiles fpc x86+Lazarus fine, then tries an x64 cross compile followed by an LCL cross compile):

Compiles along for x64 as shown here:
C:/development/fpctrunk/compiler/ppcrossx64.exe -Twin64 -Px86_64 -Ur -Xs -O2 -n -Fi../inc -Fi../x86_64
-FDC:/development/fpctrunk/bin/x86_64-win64 -FE.
-FUC:/development/fpctrunk/rtl/units/x86_64-win64 -gw -gl -dx86_64 -dRELEASE ../inc/lnfodwrf.pp
make.exe[3]: Leaving directory `C:/development/fpctrunk/rtl/win64'

.... but then it starts to try i386 output:!?!?
make.exe[1]: Entering directory `C:/development/fpctrunk' C:/development/fpcbootstrap/make.exe -C packages all make.exe[2]: Entering directory `C:/development/fpctrunk/packages' C:/development/fpcbootstrap/make.exe -C fpmkunit bootstrap make.exe[3]: Entering directory `C:/development/fpctrunk/packages/fpmkunit' C:/development/fpcbootstrap/gmkdir.exe -p units_bs/i386-win32 C:/development/fpctrunk/compiler/ppc.exe src/fpmkunit.pp -n -FuC:/development/fpctrunk/rtl/units/i386-win32
-FuC:/development/fpctrunk/packages/paszlib/units/i386-win32 -FuC:/development/fpctrunk/packages/fcl-process/units/i386-win32 -FuC:/development/fpctrunk/packages/hash/units/i386-win32 -FUunits_bs/i386-win32 -Fu../paszlib/src -Fu../hash/src -Fi../paszlib/src -Fi../fcl-process/src/win -Fu../fcl-process/src -Fi../fcl-process/src/win32 -Fi../fcl-process/src/dummy -gw -gl make.exe[3]: Leaving directory `C:/development/fpctrunk/packages/fpmkunit' C:/development/fpctrunk/compiler/ppc.exe fpmake.pp -n -FuC:/development/fpctrunk/packages/fpmkunit/units_bs/i386-win32 -FuC:/development/fpctrunk/rtl/units/i386-win32 -gw -gl Fatal: Can't find unit process used by fpmkunit
Fatal: Compilation aborted

Reinier Olislagers wrote

I wonder if the problems below are caused by wrong make files or problems in fpcup: (fpc trunk, Windows; fpcup compiles fpc x86+Lazarus fine, then tries an x64 cross compile followed by an LCL cross compile):

It's probably the same problem you had last time: the fact that "make distclean" does not work properly. In the mean time, it has been partially fixed in that running "make distclean" twice right after each other does clean everything. Try that and building then.

Jonas

Thanks Jonas - meanwhile I did implement a radical distclean alternative. You're right, I should really check again with a clean slate (i.e. clean svn checkout).

Anyway, great news about the double distclean fixing things, I can fairly easily use that in fpcup as a workaround rather than trying to duplicate things!

Thanks,
Reinier

Thanks, Jonas, changing fpcup to run make distclean 2x seemed to fix this. (Leaving some LCL x64 problems but that's another story)