Re: llvminfo woes

On 2026-07-23 10:52, Adriaan van Os via fpc-pascal wrote:
> On 22/07/2026 23:14, Jonas Maebe wrote:
>> To get a regular compiler built using an LLVM compiler, execute your 
>> make command inside the compiler directory (adding the options to OPT) 
>> and keep the resulting compiler binary (ppca64 in this case).
> 
> Thanks. That works with -O2 and produces a ppca64. With -O3 and -O4, I 
> get the results below. Adding -Swe- or -Sw- didn't help.
> 
> Assembling (pipe) aarch64/units/aarch64-darwin/nmat.ll
> Assembling (pipe) aarch64/units/aarch64-darwin/nmem.ll
> nld.pas(1271,7) Note: Call to subroutine "procedure 
> arrayconstructor_to_set(var p:tnode);" marked as inline is not inlined
> Assembling (pipe) aarch64/units/aarch64-darwin/nld.ll
> htypechk.pas(1463,12) Warning: (treated as error) Function result 
> variable does not seem to be initialized

That's a wrong error, but I cannot reproduce it. I did get a similar 
error in ncal.pas, but that one was a real mistake and I pushed a fix 
for it.

I've also pushed a fix for the issue with nested functions in TP-style 
object constructors.


Jonas
_______________________________________________
fpc-pascal maillist  -  [email protected]
https://lists.freepascal.org/cgi-bin/mailman/listinfo/fpc-pascal