[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[sc-dev] Still occasional memory corruption...
- To: sc-dev <sc-dev@xxxxxxxxxxxxxxxx>
- Subject: [sc-dev] Still occasional memory corruption...
- From: James Harkins <jamshark70@xxxxxxxxx>
- Date: Tue, 19 Nov 2013 22:42:48 +0800
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=from:to:subject:date:user-agent:mime-version:message-id :content-type:content-transfer-encoding; bh=4aQycB7KGJmw0ngsUf9IxogQ8uXikVSvZY0CeOB12Lw=; b=hxFNcEesxMCwKlkbM86MJMAyXjmLyrDNghOnkm8HxClLvLmXLQU11wJLNzPCY59AxF 2fBeYB9tgzKodyATBpVJ+0p5bfScBGChKQqi6Hv/FLVa8begSocDXxSQA12sHWWFnPmW mpIZzrnfg7vLE5mEBM29iqJ3U6zrJaOWW0J6J0agWmuRDD6kczGcSRQJJ67S5/Zi4Lzh 2v14RuylGOMWAB/7xSRQlwSFspe7BLf/FoyrAYLF6hhBO2QXcyNQDO5jYVVVFVYlTPuF Sa/2b2eSZnC6+wKwwtIQBh1y1McYxc8SboXCAjVd2xRdg370+KcPb5ijb2+cIUtlBWD3 88CA==
- List-id: SuperCollider developers mailing list <sc-devel.create.ucsb.edu>
- Reply-to: sc-dev@xxxxxxxxxxxxxxxx
- Sender: owner-sc-dev@xxxxxxxxxxxxxxxx
- User-agent: Trojita/v0.3.96-git; Qt/4.8.1; X11; Linux; Ubuntu 12.04.3 LTS
Sorry to report that I'm seeing some sporadic (very sporadic) memory
corruption problems again. It doesn't happen very often, but when things
get complicated, it does still pop up.
The latest incident was calling a pseudo-method 'addResponder' on a Proto
object. I had saved the object in a function-local variable 'proc', but
when it was time to call proc.addResponder, somehow the receiver ended up
being a TempoClock. ???? There's no possible way that could have happened
in the code. Nothing inside the function reassigns the variable, and
nothing outside the function is able to see the variable.
Unfortunately, very hard to reproduce. This happened after a few hours of
heavy use, recompiling the library a few times, but more often unloading
all the objects for the piece (lots of Protos, which include big
Environments) and re-creating them. Usually this isn't a big problem in
itself, but it added the new element of OSC control from my tablet, going
through dispatch objects and accumulating big call stacks.
Anyway, just worth noting that something somewhere in the VM internals is
still wreaking some mischief.
sc-dev mailing list
info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml