<d1b2>
<azonenberg> huh i'm not sure who wrote this strict aliasing fix
<d1b2>
<azonenberg> but i dont know how it ever worked
<d1b2>
<azonenberg> @xfol_ ok so i have a local (not pushed yet) fix for the crash but it's also parsing your file slightly wrong
<d1b2>
<azonenberg> so the import isn't usable
<d1b2>
<azonenberg> ok so, the file reports itself as being 4168 bytes in size while actually being 2336 bytes which doesnt help
<d1b2>
<azonenberg> but we ignore that field anyway
<d1b2>
<azonenberg> @xfol_ ok so this file format isnt quite matching what i am expecting to see
<d1b2>
<azonenberg> you say the actual data was acquired at 50 Ksps?
<d1b2>
<azonenberg> and overall duration of 20 ms?
<d1b2>
<azonenberg> because the "duration" field (at byte offset 0x20 in the file) is 0x5f 70 89 30 which interpreted as a little endian float32 is 1e-9
<d1b2>
<azonenberg> then the sample interval at byte 0x2c (little endian float64) is reported as 1e-12 which is 1 picosecond assuming the base unit is seconds like it's supposed to be
<d1b2>
<azonenberg> i wonder if the x and y axis units are different than i expect i'll look into this (they're reoirted as 2 and 1 respectively)
<d1b2>
<azonenberg> reported*
<d1b2>
<azonenberg> @bvernoux you have a MSO5000 right? can you have a look at this?
<_whitenotifier-4>
[scopehal] azonenberg e2b9967 - BINImportFilter: fixed broken strict aliasing fix. Fixes crash in #975, but the linked data file still doesn't import properly
<d1b2>
<azonenberg> (plus i think you wrote the original version of that filter)
<_whitenotifier-4>
[scopehal-apps] azonenberg 1dd9cac - Updated to latest scopehal
<_whitenotifier-4>
[scopehal-apps] azonenberg ec199cc - Moved PROJECT line to beginning of cmakelists. Fixes #848. Compile as C++ 20 on Windows so that vulkan-hpp pulls in the correct <version> header rather than the deprecated <ciso646> which causes warnings. Fixes #843.