Hi, On Tue, 08 Mar 2022 at 11:12, Ludovic Courtès <[email protected]> wrote:
>> valgrind: Fatal error at startup: a function redirection [...] >> valgrind: Cannot continue -- exiting now. Sorry. > > I’m late to the party, but here’s the trick you can use: Well, I am even later to the party. ;-) > --8<---------------cut here---------------start------------->8--- > $ guix build -e '(@@ (gnu packages commencement) glibc-final)' > /gnu/store/xgarlnscgbm9l04karj15xpc4nn14ak1-glibc-2.33-debug > /gnu/store/5h2w4qi9hk1qzzgi1w83220ydslinr4s-glibc-2.33 > /gnu/store/4jdghmc65q7i7ib89zmvq66l0ghf7jc4-glibc-2.33-static > $ valgrind > --extra-debuginfo-path=/gnu/store/xgarlnscgbm9l04karj15xpc4nn14ak1-glibc-2.33-debug/lib/debug > uname -o > ==16741== Memcheck, a memory error detector > ==16741== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al. > ==16741== Using Valgrind-3.18.1 and LibVEX; rerun with -h for copyright info > ==16741== Command: uname -o > ==16741== > GNU/Linux > ==16741== > ==16741== HEAP SUMMARY: > ==16741== in use at exit: 0 bytes in 0 blocks > ==16741== total heap usage: 2,813 allocs, 2,813 frees, 184,302 bytes > allocated > ==16741== > ==16741== All heap blocks were freed -- no leaks are possible > ==16741== > ==16741== For lists of detected and suppressed errors, rerun with: -s > ==16741== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0) > --8<---------------cut here---------------end--------------->8--- This trick is really nice… > I think the reason it no longer works out of the box is because we > now[*] use ‘--strip-unneeded’ instead of ‘--strip-debug’, since commit > e0f31baacc6ad30096a332b69433c85f5830bb2c. > > We’d need to double-check the hypothesis and fix it on ‘core-updates’. …but since core-updates is moving, maybe it is time for this double-check. Well, gentle reminder. :-) How could I double-check without rebuilding locally the world? Cheers, simon
