Mixxx

ASSERT on loading track to paused but moving deck

Bug #533431 reported by Sean M. Pappalardo on 2010-03-06

This bug affects 1 person

Affects		Status	Importance	Assigned to	Milestone
	Mixxx	Fix Released	High	Unassigned	Mixxx 1.8.0

Bug Description

This occurs in trunk r2358.

Steps to reproduce:
1) Remove the "- 1.0f" in ratecontrol.cpp line 363 "rate = scratchFactor - 1.0f + jogFactor + wheelFactor/10.;"
2) Build and start Mixxx
3) Load a track.
Observe Fatal abort: Fatal: []: ASSERT: "unscaled_samples_needed >= 0" in file src/engine/enginebufferscalelinear.cpp, line 140

See original description

Revision history for this message

Sean M. Pappalardo (pegasus-renegadetech) wrote on 2010-03-06:

Back trace Edit (21.4 KiB, text/plain)

description:	updated
summary:	- Crash on MP3 load with MIDI controller + ASSERT on MP3 load with MIDI controller

Revision history for this message

Phillip Whelan (pwhelan) wrote on 2010-03-08: Re: ASSERT on MP3 load with MIDI controller

This looks to me like some kind of bizarre race condition with rate. It might also be related to the low latency/lerp bug.

Revision history for this message

Sean M. Pappalardo (pegasus-renegadetech) wrote on 2010-03-08:

Whoops, this was a result of having changed something (removed the +1.0f on scratchFactor) in ratecontrol.cpp in my working copy such that rate was being set to a value <1. But that shouldn't cause a problem, right? The track would just go really slow.

Sean M. Pappalardo (pegasus-renegadetech) on 2010-03-08

summary:	- ASSERT on MP3 load with MIDI controller + ASSERT on loading track to playing deck
summary:	- ASSERT on loading track to playing deck + ASSERT on loading track to moving deck (0< rate <1)
description:	updated

Sean M. Pappalardo (pegasus-renegadetech) on 2010-03-08

summary:	- ASSERT on loading track to moving deck (0< rate <1) + ASSERT on loading track to moving deck (not in play status though)
summary:	- ASSERT on loading track to moving deck (not in play status though) + ASSERT on loading track to paused but moving deck

Revision history for this message

RJ Skerry-Ryan (rryan) wrote on 2010-06-17:

This is related to the issue we were seeing with the same assert happening when holding scratch still. For some reason when rate is 0, the EB is still processing samples. This should not be.

RJ Skerry-Ryan (rryan) on 2010-06-18