<wpwrak>
lekernel: btw, why are only PNGs shared but JPEGs aren't ? (you record the file name for PNGs only. since you explicitly handle NULL file names, this looks intentional ?)
mumptai [mumptai!~calle@brmn-4db77299.pool.mediaWays.net] has joined #milkymist
<wolfspraul>
wpwrak: the idea of adding internal 5*2 headers for USB got me thinking... are there other standard headers on PC mainboards that, if offered on M1, could allow people to plug in existing cheap stuff?
<wolfspraul>
I won't go search now, but it's something I will keep an eye open for.
<wolfspraul>
the header we are discussing now has no protection against reverse insertion. wondering how bad the damage would be if peopel inserted wrong, or how we could steer away from that
<wolfspraul>
(bright yellow connector on left side)
<wolfspraul>
that looks like 2*5 but in a smaller pitch, and embedded in this plastic to enforce correction position
kilae [kilae!~chatzilla@catv-161-018.tbwil.ch] has joined #milkymist
<GitHub14>
[flickernoise] sbourdeauducq pushed 2 new commits to master: http://git.io/WCziNg
<GitHub14>
[flickernoise/master] pixbuf: record filename for JPEG (thanks Werner for reporting) - Sebastien Bourdeauducq
<GitHub14>
[flickernoise/master] Merge branch 'master' of github.com:milkymist/flickernoise - Sebastien Bourdeauducq
cxadams [cxadams!~cxadams@li155-180.members.linode.com] has joined #milkymist
<wpwrak>
wolfspraul: the 5x2-1 header does protect against reverse insertion :) and yes, some are shrouded, for additional protection (but also possibly with additional complication)
<wpwrak>
lekernel: ah, you shouldn't have :) i;m planning to switch all this to passing FILE * to the loaders anyway, so that we don't do multiple lookups, risking races
<GitHub182>
[flickernoise] wpwrak pushed 5 new commits to master: http://git.io/tNFFEQ
<GitHub182>
[flickernoise/master] pixbuf: open and close file in pixbuf_get instead of loaders - Werner Almesberger
<GitHub182>
[flickernoise/master] pixbuf: simplify manager.c and catch strdup() failure - Werner Almesberger
<GitHub182>
[flickernoise/master] pixbuf: also compare modification time of file when looking for cache hits - Werner Almesberger
<wpwrak>
hmm, for regression testing of the pfpu itself, it would be nice to be able to run it on a PC. but i guess that's not so easy. (if we want accurate behaviour, e.g., a precise replica of rounding errors)
<wpwrak>
the second best may be to have some way to make the M1 execute a small code fragment, e.g., received form the command line, and send the results back
<wpwrak>
that may be the opportunity for using "expect" ;-))
<wpwrak>
half a lifetime ago, i think i wrote a scriptable telnet ... wish i remembered if that's still around somewhere. probably not.
wolfspra1l [wolfspra1l!~wolfsprau@p5B0AEC6F.dip.t-dialin.net] has joined #milkymist
<wpwrak>
kewl. found a gcc diagnostics error :)
lekernel [lekernel!~lekernel@f052069240.adsl.alicedsl.de] has joined #milkymist