Growl not forwarding message to Notification Center

121 views
Skip to first unread message

Kevin Walzer

unread,
Nov 2, 2013, 8:49:08 AM11/2/13
to growl-de...@googlegroups.com
Are there any known situations where Growl will not forward some
notifications to Notification Center on Mavericks?

I maintain an open-source Tcl binding to the GNTP protocol
(http://sourceforge.net/p/tk-components/code/HEAD/tree/tclgrowl/), and
while it successfully posts notifications to Growl, these are not
fowarded to Notification Center if I have that preference setting
checked. The forwarding silently fails.

Looking at a few other bindings to the GNTP protocol, I see that a
popular Python library (http://pythonhosted.org/gntp/) does seem to send
the notifications to Notification Center successfully, while a Perl
binding
(http://search.cpan.org/~mattn/Growl-GNTP-0.20/lib/Growl/GNTP.pm) does not.

Since my binding, and the Perl one, do work correctly with Growl as
designed, I am wondering where the breakdown occurs with the interaction
with Notification Center. Any suggestions?

--Kevin


--
Kevin Walzer
Code by Kevin/Mobile Code by Kevin
http://www.codebykevin.com
http://www.wtmobilesoftware.com

Daniel Siemer

unread,
Nov 2, 2013, 10:30:32 AM11/2/13
to growl-de...@googlegroups.com
Think I might have figured this out, in 2.1 we added a check for messages coming from our Growl.framework 3.0 so that we could add support for new features in the 3.0 framework (not GNTP spec, NSDistributedNotificationCenter stuff to support extra click feedback type things, as well as remotely commanding NC to fire based on rules and what not). We do a check of two headers
Origin-Software-Name
Origin-Software-Version
and the former has to be Growl.framework and the latter 3.0 or higher (We also check whether the sender was local host, but that doesn’t seem likely to factor in). However, it looks like we might generate false positives if we don’t have at least the Origin-Software-Name header, as we message a possibly nil object. This looks like there could also be a problem if you sent that you were Growl.framework but not what version you are (please don’t say you are Growl.framework if you start sending that header). Going to go file a bug on this one so that rudy and I remember to fix it so that you don’t have to send those if you don’t want to.
> --
> You received this message because you are subscribed to the Google Groups "Growl Development" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to growl-developm...@googlegroups.com.
> To post to this group, send email to growl-de...@googlegroups.com.
> Visit this group at http://groups.google.com/group/growl-development.
> For more options, visit https://groups.google.com/groups/opt_out.

Kevin Walzer

unread,
Nov 2, 2013, 8:55:18 PM11/2/13
to growl-de...@googlegroups.com
On 11/2/13, 10:30 AM, Daniel Siemer wrote:
> Think I might have figured this out, in 2.1 we added a check for messages coming from our Growl.framework 3.0 so that we could add support for new features in the 3.0 framework (not GNTP spec, NSDistributedNotificationCenter stuff to support extra click feedback type things, as well as remotely commanding NC to fire based on rules and what not). We do a check of two headers
> Origin-Software-Name
> Origin-Software-Version
> and the former has to be Growl.framework and the latter 3.0 or higher (We also check whether the sender was local host, but that doesn�t seem likely to factor in). However, it looks like we might generate false positives if we don�t have at least the Origin-Software-Name header, as we message a possibly nil object. This looks like there could also be a problem if you sent that you were Growl.framework but not what version you are (please don�t say you are Growl.framework if you start sending that header). Going to go file a bug on this one so that rudy and I remember to fix it so that you don�t have to send those if you don�t want to.

I've filed a bug on this here:

https://code.google.com/p/growl/issues/detail?id=622

Hope this helps,
Reply all
Reply to author
Forward
0 new messages