Fluid.waveform incorrectly displaying first sample

[MattS6464]

fluid.waveform appears to add an extra sample at the start and cuts off the last sample. This happens on both the audiobuffer layer and the featuresbuffer layer.

[tremblap]

Sharing the code to reproduce would be great next time @MattS6464 :) especially since it is sooooo clean, I love your Max style.

Here it is re-coded and dirty and without comments

----------begin_max5_patcher----------
663.3ocwWssbaBCD8Y3qPCO65Aws.8o7ezoiGAHbkGtMRBGmISx2dkVANtw1
XEWZyCfrDKb3b1ytH+hqiWd2ApvC8czOPNNu353.KoWvYbtiWC4PQMQ.g40P
EBxVp2Jy0jzCRX8hZJguBQJKqIOS4nJJQNvoh7gpJ0TIUHQbZ4z80qtDsURj
rt1M0rVZQ2PK7jBFincnoaPVSk.t3wUYk.Zc469FNX5gYhS9bO0vDOOzOmvg
HK9Eqc6FNsPZtZDNas+JTTBLD3+.LDr1+3Mofl0NgbfdsWcc0mVYoHsSLv7r
lGgGEEBmzPkT9FZKIuFXi+s4XEqVEeCbAup5AV45mH6oUc7l2VuS3caoHKTO
DGAyB0CWVKv2gVXqggLTx5NwsXu9g+q7Aow5y3GB9hsAoe0t.re1+NWvwWkK
qGImqGOb0rZUcGQYOl8G0LALNqCHCCteL39ChSzCIQmPaic7np9Aa46JR7cn
HszmTz7rxBCjuYZX9nfzzKPwWVzBNWzRtsnkSZ2NqpfwPGwPsRnbEg9yWXfW
NtKQ4nbq4Z7U4JPQqnZropOCnZXv+MlZpO0lz.UBG8XoJayZguF94Z9EccQ.
LRyS+.CiifTLNIc44+0596i7WGnNhTGIpiTqYb380sGO1jy7U+r3kua+UR0p
5tRigz1umMuudVR5axj9.8h+D4SHBO8Fw9vtAATzq+mLWzMvKlduFqFQuiyI
94ShQuiMcPWTbsEnoGx7HEtDHgs.ozE.Hs1guEivK.PQVPn3E.mPKvIZIxPV
fy4NAiWmz2umxEiACPnpm20w0SSWASYslowvTNcOaJdyJDtpBTpJ+T+SGn98
PhogrWSWoZiCps+YpyTjSAIzqPuaBQOwvCnkh6qt+F.TmTi4
-----------end_max5_patcher-----------

[tremblap]

investigation results:

1. adding at line 666 of waveform.js ld.data.forEach(function(s){post(s.data + '\n');}); shows that the array passed to fav is the right one...

2. pinging around line 767 of the generated fav-max.js, which is line 239 of the src/display-jsui.js, I can see that the error is in the resampling. desc[0] and desc[1] are right when printed, but the resampling is not right.

I now need to learn how to find out which definition of the sample method is used as I cannot find it anywhere to fix it...

[tremblap]

using this streamlined patch where the waveform object has 20 pixel wide for a 5 sample-long buffer makes the resampling data obvious. Printing desc[1] and its resampling max is good to understand the issue.

[tremblap]

found it. a fav.js issue. PR made there and here.