[PD] [soundfiler] - too many open files / bad file descriptor error (meaning?)
reduzent at gmail.com
Fri Feb 8 10:55:41 CET 2019
I see that your version of [conv~] uses [else/dir] internally and this
external of yours is still leaking file handles. I think I already
reported this, but can't find an open issue about it. Maybe you closed
them, despite them not being fixed.
Anyway, last time I checked, each instance of [dir] was not leaking
three file handles anymore, but only one. So there is an improvement,
but the issue is not fixed. You really need to close all file handles
that you opened before.
P.S.: I'm kinda happy to see that it bites you. This may sound mean,
but I do think that else is a great library and thus the more danger is
in such issues not being fixed. So: Thanks for the library, thanks for
fixing the issue!
On Fri, 2019-02-08 at 03:57 -0200, Alexandre Torres Porres wrote:
> Hi, I've been getting two errors involving [soundfiler], and before
> reporting an issue or sharing the patch to see what's ruining it, I'd
> like to understand and make sense of these errors, cause I just have
> no idea why they're showing up.
> First is "soundfiler_read: test.wav: Too many open files".
> Well, this is from my partitioned convolution abstraction. Inside the
> abstraction I call another abstraction in [clone]. So when I get this
> too many files error it also just prevents the abstraction from
> loading inside clone, where I get "clone: can't create subpatch
> And now things hget just crazier, cause if I try to recreate the
> asbstraction, it doesn't create in anyway, as if it got corrupted
> And if I try to save the patch, I cant', I get an error like
> "~/porres/test.pd: Bad file descriptor". All I have left to do is
> force quit Pd so I can close this file and reopen it.
> Ok, so the "too many files" is probably related to the number of
> instances I'm loading in clone, if there are too many partitions,
> this happens. But I know it shouldn't, as I was being able to load
> way more instances and I don't know how this error suddenly came
> up... none of it makes sense to me, specially what happens after this
> "too many files" error.
> probably better to mention something else separately, but I've also
> experienced [clone] acting weird, where it seemed to load old
> versions of abstractions that were already updated or deleted. This
> may be related to the fact that I cannot recreate the abstraction
> once it gets corrupted, so it keeps posting "Bad file descriptor"...
> Pd-list at lists.iem.at mailing list
> UNSUBSCRIBE and account-management ->
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 488 bytes
Desc: This is a digitally signed message part
More information about the Pd-list