IRC channel logs
2025-02-10.log
back to list of logs
<stikonas>fossy: there might be more stuff that needs double checking in live-bootstrap (possibly those early tools like script generator, etc...) <stikonas>(likely things will work since + and - are not affected) <stikonas>hmm, indeed something doens't work: configurator.c:61:Entry is not a defined symbol <stikonas>line number doesn't seem to match. Line 61 doesn't have anything... <oriansj>pointer arithmetic should just be a code smell (a bad one); but sure when I finish fuzzing M2-Planet (and fixing all of the regressions I found) <fossy>it's been a recurring bug for me for a few years that error messages don't match line numbers <oriansj>fossy: macro expansion usually; but keeping line numbers in sync after expansion is something M2-Planet should be doing correctly <fossy>hmm, i don't use macros very much... <fossy>especially multiline macros... <oriansj>M2libc's limits macros to just #ifdef and #if sort of architecture specific bits <oriansj>(and #defines where required to support the C standard) <oriansj>honestly I wouldn't even have typedef if it wasn't for the C standard FILE* which should have been called struct FILE* (But C really punted on proper type support early on) <stikonas>I'll now test live-bootstrap with your latest fix <stikonas>gtker: ok, I think thinks are looking alright, my live-bootstrap run is now running in mes stage... <stikonas>so far none, still running fine in mes-m2 stage. I guess most of the pointer arithmetic usage was in + and - operators, not in += and -= <stikonas>+= and -= are fairly recent additions anyway <stikonas>janneke: so we have partially merged pointer arithmetic support to M2-Planet <stikonas>so far just for += and -= operators, so for now mes is not really affected <oriansj>one minor note, BOOTSTRAP_MODE is for checking M2-Planet for cc_* compatibility (to catch us before we accidentially break the cc_* to M2-Planet bootstrap step)