All information is supplied in detail to re-produce,If i've not missed anything. If further information was necessary I'll do my best to provide any.
Any idea why arm-tcc would not escape on __asm__ token to process inline assembly, if this was the problem? I do know already asm processing with any other than ARCH=i386 was absent and/or incomplete, but i wanted to give arm-tcc another try regardless. On 2025-05-07 10:30, grischka wrote: > On 07.05.2025 03:05, Michael Ackermann via Tinycc-devel wrote: > > $ arm-tcc --version > > tcc version 0.9.28rc 2025-02-13 HEAD@f8bd136d* (ARM eabihf Linux) > > > > $ pwd > > musl-1.1.24-src > > > > $ /usr/bin/arm-tcc -std=c99 -nostdinc -ffreestanding \ > > -D_XOPEN_SOURCE=700 -I./arch/arm -I./arch/generic -Iobj/src/internal \ > > -I./src/include -I./src/internal -Iobj/include -I./include -O1 -Wno-error > > \ > > -static -DBROKEN_EBX_ASM -DSYSCALL_NO_TLS=1 -DCRT -c \ > > -o obj/crt/Scrt1.o crt/Scrt1.c > > > > Seems arm-tcc doesn't pick up the token for inline __asm__() I haven't the slightest clue why i386-tcc succeeds with compiling/linking entire libc and a complete Operating System distribution on my computer with i386-tcc, although some trivial patching to ~500builds, kernel fork, libc fork were necessary for this, meanwhile arm-tcc fails still. > > Hi, > > I would say that certainly an attempt to compile a complete system It's not an attempt, with i386-tcc this succeeded already, which provides a valuable baseline for any other attempt with arm/64, x86_64, and maybe riscv64. > like yours has big potential to improve tinycc, however in order > to benefit from it (rather than just produce frustration), there > are still only two options I'm afraid: > > 1) you find the problem yourself (and possibly push or propose a fix) I've provided various patches already whenever possible. > 2) you investigate the issue at least to the point where you can show > a reproducible test case. ? > > In general, if tcc produces an error message such as > > ./arch/arm/crt_arch.h:18: error: operand expected > > the purpose of that message is to help its user in the first place > rather than to be pasted into an email and to be sent to the devel-list. > How should anyone else be able to find out more easily than you what is > at line 18 of ./arch/arm/crt_arch.h which is a file on your computer. It is musl-libc 1.1.24 the problem with arm-tcc was spotted with, not any arbitrary file on my computer, and i have provided detailed information how to reproduce the problem. Does that mean the problem was not re-producible elsewhere? Then I myself must have made a mistake, that's good to know then. Thank You. > > -- grischka > -- Michael Ackermann
signature.asc
Description: Digital signature
_______________________________________________ Tinycc-devel mailing list Tinycc-devel@nongnu.org https://lists.nongnu.org/mailman/listinfo/tinycc-devel
