IRC channel logs

<dddddd>That gives the same results as with M1.scm, cleanly. But this is gcc. Now M2-Planet is another history.

<OriansJ>M2-Planet can use just plain unsigned

<OriansJ>eg unsigned i = value;

<dddddd>Currently, M2-Planet >> behaviour is the same for unsigned.

<dddddd>... for every arch.

<dddddd>But, at least, we have a fix/workaround for M1-macro.c

<dddddd>- int i = value;

<dddddd>+ unsigned i = value;

<dddddd>Which, in retrospect... seems obvious as shift for signed is arithmetic, not logical :/

<dddddd>And... for the record, the "char* s" parameter is "output".

<dddddd>The original function, for the curious readers:

<dddddd>int stringify(char* s, int digits, int divisor, int value, int shift)

<dddddd> int i = value;

<dddddd> if(digits > 1)

<dddddd> i = stringify(s+1, (digits - 1), divisor, value, shift);

<dddddd> s[0] = hex2char(i & (divisor - 1));

<dddddd> return (i >> shift);

<dddddd>For octal, divisor is 8, shift 3.

<fossy>I am 40% done porting coreutils 5.0 to tinycc

<fossy>reasonably simple task

<fossy>and should make the bootstrap a bit (lot) faster for tinycc to just past gcc

<damo22> http://paste.debian.net/plain/1140050 a small patch on a hurd system got me this far

<damo22>i needed `-o $o` because .S.o dont get replaced as .c -> .o

<damo22>not sure why its missing itoa and eputs

<damo22>`-L mescc-lib/$mes_cpu-mes` seemed to be missing

<fossy>whats wrong damo22

<damo22>just trying to build mes on hurd

<damo22>using wip-hurd branch

<damo22>there seems to be two sets of build scripts, configure and configure.sh

<damo22>i used .sh variant

<Hagfish>coreutils for tcc? nice! i'm guessing you can't make that a configurable option that upstream would accept

<OriansJ>fossy: interesting approach. It gives me an idea.

***ChanServ sets mode: +o rekado_

***rekado_ is now known as rekado

<fossy>OriansJ: guile or coreutils

2020-04-13.log