I reported one case already.
https://bugs.freepascal.org/view.php?id=35598

Compiling 42090 for win32 with the options
-gl -O4 -Or
fails with the warning/error: defutil.pas(1606,7) Warning: Function result variable does not seem to initialized

Just run into the next error. Compiling with -gl -gw -godwarfsets -O-1 -dTEST_WIN32_SEH

c:/FPC/fpc_3.3.1/source/compiler/ppc1.exe -XX -CX -Ur -Xs -O2 -n -O2 -Fui386 -Fusystems -Fuc:/FPC/fpc_3.3.1/source/rtl/units/i386-win32 -Fii386 -FEi386/bin/i386-win32 -FUi386/units/i386-win32 -dRELEASE -gl -gw -godwarfsets -O-1 -dTEST_WIN32_SEH    -di386 -dGDB -dBROWSERLOG -Fux86 -Sew pp.pas
n386flw.pas(142,21) Warning: Case statement does not handle all possible cases
n386flw.pas(153,21) Warning: Case statement does not handle all possible cases
n386flw.pas(675) Fatal: There were 2 errors compiling module, stopping Fatal: Compilation aborted

Op 2019-05-18 om 01:47 schreef Martin:

And if you fix it, it will just pass on to the next one (in ngen* iirc)

It's a known issue,  please pass ALLOW_WARNINGS=1  to make to avoid it for now.

Marco van de Voort wrote

n386flw.pas(142,21) Warning: Case statement does not handle all possible cases

What's this warning all about? There are far too many situations where you may not want to handle all possible cases and the implicit else block just goes to the first line after the case block - e.g. the peephole optimizer.  This really doesn't feel like a warning, more like a aka. Kit

---
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

It's a known issue,  please pass ALLOW_WARNINGS=1  to make to avoid it for now.

OK, will do.

I reported it, because a month ago it was compiling fine (with the same settings)

Op 2019-05-18 om 12:31 schreef Martin Frb:

A ticket is always a good thing as a reminder that it is not fixed. And it provides a place to document the workaround.