Backport LocalOut fix into 3.8.1?

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

Backport LocalOut fix into 3.8.1?

jamshark70-2
Hi,

I see that a 3.8.1 release is going out.

Literally just today, I just discovered that the 3.8.0 package that I recently downloaded for Windows includes a bug from September 2016 that breaks LocalIn/LocalOut. The fix was committed this past January (remarkable in itself that LocalOut was broken for 15 months and nobody noticed, but anyway).

Bug: https://github.com/supercollider/supercollider/commit/f86acb097ee295b59403727601868da513d96a63

Fix:
https://github.com/supercollider/supercollider/commit/8d634d806707881f89423c279cec659b5be99c1a

Shouldn't we backport this fix? I mean, I know we can't backport everything but this is kind of a wtf bug.

https://github.com/supercollider/supercollider/blob/3.8/SCClassLibrary/Common/Audio/InOut.sc#L220

That's the offending line, still incorrect at the head of the 3.8 branch.

I have a quark that I want to publish in the next couple of months, with some default instruments for my live coding framework. One of them uses LocalOut. If this bug isn't fixed, then I will have to document that the live coding dialect is backward compatible with 3.8, but the instruments are valid only for 3.9. That's a weird thing to have to document.

hjh


_______________________________________________
sc-dev mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-dev/
search: http://www.listarc.bham.ac.uk/lists/sc-dev/search/
Reply | Threaded
Open this post in threaded view
|

Re: Backport LocalOut fix into 3.8.1?

yvan.sc
> Literally just today, I just discovered that the 3.8.0 package that I recently downloaded for Windows includes a bug from September 2016 that breaks LocalIn/LocalOut. The fix was committed this past January (remarkable in itself that LocalOut was broken for 15 months and nobody noticed, but anyway).
>
> Bug: https://github.com/supercollider/supercollider/commit/f86acb097ee295b59403727601868da513d96a63
>
> Fix:
> https://github.com/supercollider/supercollider/commit/8d634d806707881f89423c279cec659b5be99c1a
>
> Shouldn't we backport this fix? I mean, I know we can't backport everything but this is kind of a wtf bug.

+1 !!

_______________________________________________
sc-dev mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-dev/
search: http://www.listarc.bham.ac.uk/lists/sc-dev/search/
Reply | Threaded
Open this post in threaded view
|

Re: Backport LocalOut fix into 3.8.1?

jamshark70-2
In reply to this post by jamshark70-2
---- On Wed, 20 Dec 2017 14:56:03 +0800 <[hidden email]> wrote ----
> Literally just today, I just discovered that the 3.8.0 package that I recently downloaded for Windows includes a bug from September 2016 that breaks LocalIn/LocalOut. The fix was committed this past January (remarkable in itself that LocalOut was broken for 15 months and nobody noticed, but anyway).

Possibly it affects only LocalOut.kr with a single channel. That might be why no one noticed for a long time.

Anyway, it's a pointless gotcha and the fix is known.

hjh


_______________________________________________
sc-dev mailing list

info (subscription, etc.): http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
archive: http://www.listarc.bham.ac.uk/marchives/sc-dev/
search: http://www.listarc.bham.ac.uk/lists/sc-dev/search/
Reply | Threaded
Open this post in threaded view
|

Re: Backport LocalOut fix into 3.8.1?

brianlheim
Hey James,

Thanks for the suggestion. We just intended 3.8.1 to be a super-minor release so that the Arch Linux maintainer can have a package that doesn't require patching (as per previous dev meeting minutes). But I think what you're suggesting isn't unreasonable at all. The only thing is that this would require us to add binaries for the release, whereas before it could have been source-only.

Regards,
Brian

On Wed, Dec 20, 2017 at 8:51 AM, <[hidden email]> wrote:
---- On Wed, 20 Dec 2017 14:56:03 +0800 <[hidden email]> wrote ----
> Literally just today, I just discovered that the 3.8.0 package that I recently downloaded for Windows includes a bug from September 2016 that breaks LocalIn/LocalOut. The fix was committed this past January (remarkable in itself that LocalOut was broken for 15 months and nobody noticed, but anyway).

Possibly it affects only LocalOut.kr with a single channel. That might be why no one noticed for a long time.

Anyway, it's a pointless gotcha and the fix is known.

Reply | Threaded
Open this post in threaded view
|

Re: Backport LocalOut fix into 3.8.1?

jamshark70-2

On December 22, 2017 04:23:48 [hidden email] wrote:

> Thanks for the suggestion. We just intended 3.8.1 to be a super-minor
> release so that the Arch Linux maintainer can have a package that doesn't
> require patching (as per previous dev meeting minutes). But I think what
> you're suggesting isn't unreasonable at all. The only thing is that this
> would require us to add binaries for the release, whereas before it could
> have been source-only.


After thinking about it some more, I guess it's not really critical. To me, it depends on, for instance, Linux packaging. If we expect popular distros to move up to 3.9 fairly quickly (ha!) and skip over 3.8.1, then I'd say don't bother backporting. But, if major distros would ship 3.8.1 and wait on 3.9, then it's worth including.

I don't know what's likely to happen with that. So I'll leave it to the community's judgment.

hjh

Sent with AquaMail for Android
http://www.aqua-mail.com

Reply | Threaded
Open this post in threaded view
|

Re: Backport LocalOut fix into 3.8.1?

contact
In reply to this post by brianlheim
IMHO regressions should always be fixed in the next possible minor releases.

best, .h.h.


On 21/12/17 21:23, [hidden email] wrote:

> Hey James,
>
> Thanks for the suggestion. We just intended 3.8.1 to be a super-minor
> release so that the Arch Linux maintainer can have a package that
> doesn't require patching (as per previous dev meeting minutes). But I
> think what you're suggesting isn't unreasonable at all. The only thing
> is that this would require us to add binaries for the release, whereas
> before it could have been source-only.
>
> Regards,
> Brian
>
> On Wed, Dec 20, 2017 at 8:51 AM, <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     ---- On Wed, 20 Dec 2017 14:56:03 +0800 <[hidden email]
>     <mailto:[hidden email]>> wrote ----
>     > Literally just today, I just discovered that the 3.8.0 package that I recently downloaded for Windows includes a bug from September 2016 that breaks LocalIn/LocalOut. The fix was committed this past January (remarkable in itself that LocalOut was broken for 15 months and nobody noticed, but anyway).
>
>     Possibly it affects only LocalOut.kr with a single channel. That
>     might be why no one noticed for a long time.
>
>     Anyway, it's a pointless gotcha and the fix is known.
>
>     hjh
>
>
>     _______________________________________________
>     sc-dev mailing list
>
>     info (subscription, etc.):
>     http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
>     <http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx>
>     archive: http://www.listarc.bham.ac.uk/marchives/sc-dev/
>     <http://www.listarc.bham.ac.uk/marchives/sc-dev/>
>     search: http://www.listarc.bham.ac.uk/lists/sc-dev/search/
>     <http://www.listarc.bham.ac.uk/lists/sc-dev/search/>
>
>


signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Backport LocalOut fix into 3.8.1?

brianlheim
Actually, this brings up a good question about patch release support. Do we currently have a section on this anywhere in our wiki? I feel like previously we have agreed that we ought to only guarantee patch releases for the current minor release, but not prohibit patch releases on older minor releases. That's the policy that would make sense to me anyway.

Brian

On Fri, Dec 22, 2017 at 4:49 AM <[hidden email]> wrote:
IMHO regressions should always be fixed in the next possible minor releases.

best, .h.h.


On 21/12/17 21:23, [hidden email] wrote:
> Hey James,
>
> Thanks for the suggestion. We just intended 3.8.1 to be a super-minor
> release so that the Arch Linux maintainer can have a package that
> doesn't require patching (as per previous dev meeting minutes). But I
> think what you're suggesting isn't unreasonable at all. The only thing
> is that this would require us to add binaries for the release, whereas
> before it could have been source-only.
>
> Regards,
> Brian
>
> On Wed, Dec 20, 2017 at 8:51 AM, <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     ---- On Wed, 20 Dec 2017 14:56:03 +0800 <[hidden email]
>     <mailto:[hidden email]>> wrote ----
>     > Literally just today, I just discovered that the 3.8.0 package that I recently downloaded for Windows includes a bug from September 2016 that breaks LocalIn/LocalOut. The fix was committed this past January (remarkable in itself that LocalOut was broken for 15 months and nobody noticed, but anyway).
>
>     Possibly it affects only LocalOut.kr with a single channel. That
>     might be why no one noticed for a long time.
>
>     Anyway, it's a pointless gotcha and the fix is known.
>
>     hjh
>
>
>     _______________________________________________
>     sc-dev mailing list
>
>     info (subscription, etc.):
>     http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx
>     <http://www.birmingham.ac.uk/facilities/ea-studios/research/supercollider/mailinglist.aspx>
>     archive: http://www.listarc.bham.ac.uk/marchives/sc-dev/
>     <http://www.listarc.bham.ac.uk/marchives/sc-dev/>
>     search: http://www.listarc.bham.ac.uk/lists/sc-dev/search/
>     <http://www.listarc.bham.ac.uk/lists/sc-dev/search/>
>
>