hello,
It seems that the pd-osc package in Ubuntu Utopic is buggy. Values are not parsed correctly and I got for example : "gatom: no method for '' "
But I noticed that the up-to-date svn does not suffer from this. So I guess the package is outdated.
Where should I report this ? Here ? On the SF bug tracker ? On Launchpad ? ... ?
cheers
antoine
-- do it yourself http://antoine.villeret.free.fr
On 11/05/2014 05:42 PM, Antoine Villeret wrote:
hello,
It seems that the pd-osc package in Ubuntu Utopic is buggy. Values are not parsed correctly and I got for example : "gatom: no method for '' "
But I noticed that the up-to-date svn does not suffer from this. So I guess the package is outdated.
Where should I report this ? Here ? On the SF bug tracker ? On Launchpad ? ... ?
ideally report the bug to *Debian* (as ubuntu mostly takes the packages from Debian (e.g. in this case)).
something like: $ reportbug pd-osc
fgmdsrt IOhannes
I've done it on a debian wheezy system on Raspberry Pi with reportbug but I don't know if it's well arrived since I didn't receive an e-mail with bug Id...
I notice that the problem occurs only when parsing paths with more than one "/"
see patches attached, it uses iemnet too
a
-- do it yourself http://antoine.villeret.free.fr
2014-11-05 19:38 GMT+01:00 IOhannes m zmölnig zmoelnig@iem.at:
On 11/05/2014 05:42 PM, Antoine Villeret wrote:
hello,
It seems that the pd-osc package in Ubuntu Utopic is buggy. Values are not parsed correctly and I got for example : "gatom: no method for '' "
But I noticed that the up-to-date svn does not suffer from this. So I guess the package is outdated.
Where should I report this ? Here ? On the SF bug tracker ? On Launchpad ? ... ?
ideally report the bug to *Debian* (as ubuntu mostly takes the packages from Debian (e.g. in this case)).
something like: $ reportbug pd-osc
fgmdsrt IOhannes
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
On 11/05/2014 11:10 PM, Antoine Villeret wrote:
I've done it on a debian wheezy system on Raspberry Pi with reportbug but I don't know if it's well arrived since I didn't receive an e-mail with bug Id...
you should have. if not something went wrong... you could also create a bug-report directly (via email) using the instructions found in [1].
btw, i don't get a problem with your patch, though i replaced the udp-part with a simple [s]/[r] pair and ran both patches in the same Pd-instance.
gfmadsr IOhannes
2014-11-06 10:26 GMT+01:00 IOhannes m zmölnig zmoelnig@iem.at:
On 11/05/2014 11:10 PM, Antoine Villeret wrote:
I've done it on a debian wheezy system on Raspberry Pi with reportbug but I don't know if it's well arrived since I didn't receive an e-mail
with
bug Id...
you should have. if not something went wrong... you could also create a bug-report directly (via email) using the instructions found in [1].
yes I saw that too, I will report it there
btw, i don't get a problem with your patch, though i replaced the udp-part with a simple [s]/[r] pair and ran both patches in the same Pd-instance.
here is a minimal patch that remonstrate the issue [print bug] prints the right float to console but the number box doesn't understand it...
this is of course with pd-osc-1.0-2 from wheezy current SVN revision works well
a
gfmadsr
IOhannes
[1] https://www.debian.org/Bugs/Reporting
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
On 11/06/2014 11:21 AM, Antoine Villeret wrote:
2014-11-06 10:26 GMT+01:00 IOhannes m zmölnig zmoelnig@iem.at:
here is a minimal patch that remonstrate the issue [print bug] prints the right float to console but the number box doesn't understand it...
this is of course with pd-osc-1.0-2 from wheezy current SVN revision works well
ah i see and can reproduce the problem (my last email used the wrong version of [routeOSC], so forget that).
afair, [routeOSC] used to only be able to route a *single* element from the OSC-address. so [routeOSC /a/b] was an illegal string. at least that was my understanding (as it didn't work).
i checked the actual output of the buggy example, and it really sends a message with an empty selector ("") followed by the number you were expecting. so putting an [unpack s f] after the [routeOSC] will give the number on the second outlet.
of course this is really a bug.
it's a bummer that Debian/jessie was frozen yesterday, which means that a minor bug like this won't get fixed for jessie :-(
gfmsdr IOhannes
On 11/06/2014 05:53 PM, IOhannes m zmölnig wrote:
On 11/06/2014 11:21 AM, Antoine Villeret wrote:
2014-11-06 10:26 GMT+01:00 IOhannes m zmölnig zmoelnig@iem.at:
here is a minimal patch that remonstrate the issue [print bug] prints the right float to console but the number box doesn't understand it...
this is of course with pd-osc-1.0-2 from wheezy current SVN revision works well
ah i see and can reproduce the problem (my last email used the wrong version of [routeOSC], so forget that).
afair, [routeOSC] used to only be able to route a *single* element from the OSC-address. so [routeOSC /a/b] was an illegal string. at least that was my understanding (as it didn't work).
i checked the actual output of the buggy example, and it really sends a message with an empty selector ("") followed by the number you were expecting. so putting an [unpack s f] after the [routeOSC] will give the number on the second outlet.
of course this is really a bug.
i think it was reported on sf as bug#3426523 (in the old counting. in the new counting it is [766]) and consequently fixed.
fgmsdr IOhannes
On Thu, 2014-11-06 at 17:53 +0100, IOhannes m zmölnig wrote:
On 11/06/2014 11:21 AM, Antoine Villeret wrote:
2014-11-06 10:26 GMT+01:00 IOhannes m zmölnig zmoelnig@iem.at:
here is a minimal patch that remonstrate the issue [print bug] prints the right float to console but the number box doesn't understand it...
this is of course with pd-osc-1.0-2 from wheezy current SVN revision works well
ah i see and can reproduce the problem (my last email used the wrong version of [routeOSC], so forget that).
afair, [routeOSC] used to only be able to route a *single* element from the OSC-address. so [routeOSC /a/b] was an illegal string. at least that was my understanding (as it didn't work).
i checked the actual output of the buggy example, and it really sends a message with an empty selector ("") followed by the number you were expecting. so putting an [unpack s f] after the [routeOSC] will give the number on the second outlet.
of course this is really a bug.
it's a bummer that Debian/jessie was frozen yesterday, which means that a minor bug like this won't get fixed for jessie :-(
That is really unlucky. Thanks for reporting this, Antoine. I was surprised to read about that bug, since I've been working a lot with OSC classes and never triggered it. I've used [routeOSC] in so many ways, but obviously not this :-/
Roman
yep, but unfortunately this is not the only one I found another bug in [packOSC], see [sf#1167]
[sf#1167] : https://sourceforge.net/p/pure-data/bugs/1167/
do it yourself http://antoine.villeret.free.fr
2014-11-07 9:48 GMT+01:00 Roman Haefeli reduzent@gmail.com:
On Thu, 2014-11-06 at 17:53 +0100, IOhannes m zmölnig wrote:
On 11/06/2014 11:21 AM, Antoine Villeret wrote:
2014-11-06 10:26 GMT+01:00 IOhannes m zmölnig zmoelnig@iem.at:
here is a minimal patch that remonstrate the issue [print bug] prints the right float to console but the number box doesn't understand it...
this is of course with pd-osc-1.0-2 from wheezy current SVN revision works well
ah i see and can reproduce the problem (my last email used the wrong version of [routeOSC], so forget that).
afair, [routeOSC] used to only be able to route a *single* element from the OSC-address. so [routeOSC /a/b] was an illegal string. at least that was my understanding (as it didn't work).
i checked the actual output of the buggy example, and it really sends a message with an empty selector ("") followed by the number you were expecting. so putting an [unpack s f] after the [routeOSC] will give the number on the second outlet.
of course this is really a bug.
it's a bummer that Debian/jessie was frozen yesterday, which means that a minor bug like this won't get fixed for jessie :-(
That is really unlucky. Thanks for reporting this, Antoine. I was surprised to read about that bug, since I've been working a lot with OSC classes and never triggered it. I've used [routeOSC] in so many ways, but obviously not this :-/
Roman
Pd-list@lists.iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
-
Antoine Villeret -
IOhannes m zmölnig -
Roman Haefeli