As Georg-Johann Lay wrote: > Presumably, it goes like follows: avr-libc detects from the version of > avr-gcc that it does not add -latmega128 and hence skips building > libatmega128.a, hence no such .a in the avr-libc build tree.
Yep. However, Pitchumani added (in r2520) code to runtest.sh that always assumes the current multilib directory layout, thus the $AVRDIR/avr/lib/$multilibdir/$2/lib$2.a is always added to the linker line. This, in turn, is needed since runtest.sh does not want to rely on any installed library version, but instead allows for testing inside the current build environment, so it uses -nostdlib, and adds all the libraries explicitly. I'm not sure whether it really makes much sense to put any kind of effort into this to allow being compiled on older compiler versions. I just fixed a few errors in the testsuite where code has been causing warnings or compilation errors. Being still at GCC 5.3.0 here, all the scanf tests stumble across a GCC ICE: Simulate: scanf/sscanf_flt-f1.c /scanf_flt atmega128 ... scanf/sscanf_flt-f1.c: In function 'main': scanf/sscanf_flt-f1.c:162:1: internal compiler error: in push_reload, at reload.c:1380 } ^ no stack trace because unwind library not available Please submit a full bug report, with preprocessed source if appropriate. See <http://gcc.gnu.org/bugs.html> for instructions. *** compile failed That might perhaps already be fixed in more recent GCC versions. I see a few regression errors:: Simulate: regression/bug-27242.c atmega128 ... *** simulate failed: 8 Simulate: regression/bug-27242.c at90s8515 ... *** simulate failed: 8 Simulate: stdlib/malloc-7.c atmega128 ... *** simulate failed: 92 Simulate: stdlib/malloc-7.c at90s8515 ... *** simulate failed: 92 Simulate: stdlib/realloc-3.c atmega128 ... *** simulate failed: 2 Simulate: stdlib/realloc-3.c at90s8515 ... *** simulate failed: 2 Simulate: other/realloc-01.c atmega128 ... *** simulate failed: 2 Simulate: other/realloc-01.c at90s8515 ... *** simulate failed: 2 These are all about malloc/realloc. They have to be investigated. Interesting enough, all the previously known issues in known-to-fail.txt no longer appear in the list. Finally, I had to drop AT90S4414 from the "full MCU" list since modern simulavr no longer supports it, and replaced AT90S2313 by ATtiny2313. This yields a few new simulation errors: Simulate: avr/eeprom-1.c attiny2313 ... *** simulate failed: 70 Simulate: avr/eeprom-2.c attiny2313 ... *** simulate failed: 46 Simulate: avr/eeprom-3.c attiny2313 ... *** simulate failed: 64 Not sure what's wrong here ... -- cheers, Joerg .-.-. --... ...-- -.. . DL8DTL http://www.sax.de/~joerg/ Never trust an operating system you don't have sources for. ;-) _______________________________________________ AVR-libc-dev mailing list AVR-libc-dev@nongnu.org https://lists.nongnu.org/mailman/listinfo/avr-libc-dev
