In the meantime, I’ll keep a 4.6.4 version for upload and a new one for editing.
I made the test with the options « raw data from libdivecomputer » and « libdivecomputer log ». No crash with either version, because of the « raw data » option, I guess. But crash is back when unchecking the option. « Old » version (the one working): Subsurface: v4.6.4-957-g56e755b7119b, built with libdivecomputer v0.6.0-devel-Subsurface-branch (07f5777c71b17bf1dd4dfecea6fb2c217756a647) « 4.7.2 » : Subsurface: v4.7.1, built with libdivecomputer v0.6.0-devel-Subsurface-branch (07f5777c71b17bf1dd4dfecea6fb2c217756a647) So it seems to be the same libdivecomputer version... Philippe > Le 25 oct. 2017 à 07:08, Linus Torvalds <[email protected]> a > écrit : > > On Tue, Oct 24, 2017 at 9:47 PM, Philippe Massart > <[email protected]> wrote: >> I tested previous versions: 4.6.4-1031 crashes the same way but 4.64-957 >> seems to work on both accounts. > > Hmm. There are no changes to core/libdivecomputer.c in that range: > those versions should be commits > > e4a9787c5 dive.h: add handling of NULL in get_dive_dc() > > to > > 56e755b71 Use lrint() for all degrees_t related rounding > > but there might be a libdivecomputer update in between. Sadly, the > version of libdivecomputer isn't reported. > > I don't see anything recent wrt the Mares Puck, though. There was some > cleanup back in April, but that's just about it. > > Very odd. > > Linus _______________________________________________ subsurface mailing list [email protected] http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
