Fluid.hpss~ crash when changing attributes

Hello,

I was able to make fluid.hpss~ crash max. It happened when I changed my hsize and psize in that order. I’ll try to recreate it if I can but it seemed fairly egregious so I thought I would post it anyway.

https://hastebin.com/gucuhaheyu

Okay I made it happen again. It seems to be when I change the psize.

https://hastebin.com/upegifaqoy

The two crashes are very different, but it looks like there might be a memory stamping issue here.

Yes, I noticed there were different reports, but can guarantee that I attempted to replicate them doing the exact same thing.

  1. Load patch
  2. Change hsize
  3. Change psize
  4. Crash

will check with the current refactor-review branch, it might have been sorted since a LOT changed under the hood. Will be back online with an answer in 45m

edit: I cannot reproduce whilst vigorously wiggling the parameters… but there are info missing in your report: is audio on, what dsp options, is audio running through the patch, are you on Alpha03 (which I presume you are, since they are the latest public), etc. Give me a bit more and I’ll try to reproduce a bit more…

sigvs: 64
iovs: 512
overdrive: on
interupt: off


----------begin_max5_patcher----------
1693.3ocyZs0aaaCE9YmeEBB8g0szTdWj6otsG5KanOzGFFJFBjsocTmtXHI
2jrhle6i7PIEIeQV1VtYAHNTRj9b99NWIU95US7ml8ftv26m89j2jIe8pISf
aYuwjpqm3mD9vr3vBXZ9o56yl9Y+qcOpT+PIb6u3kWdycqJJtctNN7w5mmst
LVWV93JsSF99d+c0iVEVN6tnzk2lqmU5dpPvtAcsGUA+QpreRH2fZVS55jnT
yWHnJ3paFMGz.iV8Fpv+4Y5jsap1a9sqtx9w0CDnKhyLeI6GHW64OMLcY+.B
EbCuAPbzV.ZQVdRHLYwPgH2uQd4gI5Rc9s5zvowfhg1I7Im.72ictHLYUQYV
RwS6kYJhVlFFa4GCEZAWODDmo.BhJsTSPvwawo9iFjSzEEgK0agYIVQNMGZl
h.NzN3sC6eG3Q1A7HimC8drny0Kz4wY2eZHTJrXhwA7wNdyG5hiOiO3bHPce
36fQwNLhc1P4QiQh7hiwk5xvxxbu6xVUD8u5dSZA+1GbIbwyAixiOjjrmjvz
wGu2Gkd93EKZiW9wiW92M7d24iVYKvFb7ArX4EGrkdSO8XUtfXKnPDX.ZGO.
IpQLZccxTc9Y1BgxkwAELjRH6BPAunMLneXUt2O7pHr2OY+Cw6sduJh9Zu23
g8ds2OZu3zp7vITfKr8VIBDDghSwDisGqNfUmtC2Z9+ur50f6js5L12Iq9X3
hiQPL6oC1W3dh0oyexakNeleOUUwT.jN3IN91H1SUUzXii6ByS5AGXni1JXD
PNZXvGOXLKKIQmVtENrH3sGvbvXrVAXB2VMQGiO2E2bj+jcO0QoGZiV81aC4
4VWYz9MVncT63h245h30QygCN3Iu2AMu5Qwdu6tlQUs34wwDysc82ZaQ06cK
VT5t.wjCXynaM5v7FA41cC43cywGtMI36xONJcyCfAzG686RmEYqymUiq5zd
dOqRy0EkQogkQYosljselVSJKetIc8dsgipnsEUMU42RznySzrgJZyjNKIYO
agV5+tkj8Hm1E+h9NHZ9kvzZ2R9g8ptLndPh9kC01M+rKQSNKQuQD5EzglHG
fjpYhySRAuXolFjnA5jL19uaDUrGT2U+RhluJKJsrpF.QvuA1CCTxIvc9wvU
mqEQLfLJ3dUMkpkpQvillMjZIaTvYCUCGzl0DDn4zQS2NOVCIuLr1P7xGgHY
7PxYrQM4Mn.UaiCmRFKJ.Tsyx3fcaydzMN3g3RC03G8ZnCSzjc2Y14k4EO.y
gXLbIG.B2tboqc6vUq9hNunZxfHL6S4yY.7kWCWFk5tD1Pjet9KQ0yGdCM9g
4lsKTZ1qv5bPs7eP31MneRlgHSWGUwkFvYDYQ4iwa1heZXhao+x54QYerLrb
cws+gNcsaaC1WJS353xtDwzkKhhimkE6TuIMOwreppc736d50U2tYtexCcCQ
wvXogVtghoADALxLfyaY1pVCtdQLNSgH1oxDj.FGFIITN0NBswxHOKKDV4j.
RIQL2HysvFY0dYgoKcmRCIvlzt5tqxyVkkWaRMKU0L+0kYKyCmGY23ecvhyD
6na2g.YdJP6.mTwoIFBMpdMs86ZLF+95YgmqAnQ41iMXf7RWqfrg6kAHtT5F
IYzctrFqffaDhctB6OAtQtuoKhQnJhrChYAbLETchhJUtQbLRn5nD9gylY91
5rTJw37ALDqFHV2U1Fqb5xNqRgX3.XYFowB.TKYDU.HZtnYc5XcxlhLPZ3IG
QivRoajQ5DRWYtHKsr1swjw4CqzodeLLsv6i5jnoYwye9zEJLxYlkE6JHhgA
.HgEADWvjDgvrtxwdNIv5tMJ0l2R23HQUDrvQl0Qnj1dDiR.w+TDGYxn8m1z
c8EXzBZv6coyGijtjD9fvdDS8nF6vGxEpzPVAUQc60+YC023vbXaALu5OFW.
Cmwz.YdSrlIuLQ0NNw5Tv33wVs57eJwXmqrUVu5rVaOpmrdARa8oq24nsW1K
VEqw0qpazwvrhtykbvdWftzpe.CKQUiomk6Hh+03055iOcvg4Tbff4Jv0nfJ
6svW.ELO69ziVCeVu3MwlzKjF9aOFd7JHhXJW5RRFPoDWzFAiwzwWAeetVeB
ZHvdT6OtBm11hFek6uzwwY22u1Yq+C+uXz71B3JDR1W2LRSWOAtbQXIFhq4F
FmOp..daru2z+wQStVWQDX94lhkbPSwsy+MJJXtqgBqFN37NBlPPUUtiAUMz
YHNEs29.u..xsER3cZYQTwpP29RgW80Ue6p+CFP.xH.
-----------end_max5_patcher-----------

better… I can ever say almost good: which version of Max, what OS, confirm Alpha03 use…

I’m unable to reproduce, which is most probably a good news, since the algo change, is now faster… the bad news is that the name of attributes also changed…

Sorry, comprehension before 12 (:

I’m on 10.14.3 OS, 8.0.3 Max and Alpha03

ok then I think we sorted it but we’ll need to try on Max8.

Is it possibly an issue with getattr trying to snap up the attributes as they change? I constantly run into issues with getattr, but this is the easiest way for me to keep my delay compensation variable updated across the patch. I suppose I could force it with message $1 and extract it that way if getattr is indeed the problem.

everything is dynamic now, you can change everything in realtime, and there is none of these issues, either at instantiation or wiggling them, so no worries there. :wink:

I look forward to new alpha (^: