We are pleased to announce the release of Spark 2.5.0. Spark 2.5.0 is a major milestone for the project that includes lots of new features, including Jingle VOIP support and a new look and feel. We have also been working very hard to improve the stability, performance, and startup times. For those Mac users, we have been working extra hard getting Spark “up to snuff” with our windows client release. We hope you notice :).
Btw. auto-update (Help -> Check for Updates) does not recognizes 2.5.0 as new version if updating from beta releases through Wildfire Enterprise ‘‘Spark Versions’’ facility.
I just hope all those presence related issues we are experiencing here with Spark have now been fixed, or did they?
The feature to show the status of offline users (SPARK-2) is not working for us. I can see in the traffic window that the status is being received, but nothing is displayed in the list or the popup vcard. Is there an option or something that needs to be changed to make this work?
Is this some sort of joke? Because I don’'t find it funny. Why do you release 2.5 without supporting jingle in Linux?
Are you following the footsteps of Skype? Or maybe Microsoft?
Heh. We usually try to avoid being an evil empire. In all seriousness, the choice came down to releasing now or waiting until some unforseen point in the future until we could stabilize the Linux audio support. Given our very limited resources, we choose the former, especially given all the awesome stuff in 2.5. A few things we’'ve started doing:
We massively improved Spark on Mac for 2.5.0. I think that demonstrates our commitment to cross-platform.
We’'re now doing specific testing of Spark on Linux (a huge improvement in itself). You can see the list of bugs in the issue tracker (and vote for them!)
Finally, the more feedback we can get from Linux users that they care about this stuff. Having a few paying Linux customers would never hurt as well.
Are you still uploading? There is still no Online MSI release at Downloads page.
UPD: so online msi setup is upladed now.
There is some issue with the changelogs when you are going from the Betas to Final releases. You just combine all beta changelogs, so now how can i know what exactly was fixed/added from Beta 4 to Final? This changelog is really huge, so it will take some time to check is it have some new issues fixed/added.
We’‘re moving to a release train model for Spark, which means more rapid, smaller releases (one every three weeks when the train gets fully going). So far, we know that SSO won’‘t be in 2.5.1 or 2.5.2, but we’'re doing a lot of background work on it now, and may be able to ship it in June. Unfortunately, SSO is extremely complicated on both the client and server side (at least for us newbs).
Guys, I appreciate it’‘s complicated technology and I won’‘t lie, this disappoints me but I’‘m more annoyed by the fact you weren’'t straight with us when asked a direct question last week!
That’‘s how your end users lose faith. If you’‘re now talking June for this functionality then there’'s no way on earth it was ever going to make it into the 2.5 release. Why not just be straight with us from the outset?
I’‘ve now got to go to my boss and tell him he’‘s not getting SSO as promised and he’'s likely going to tell me to scrap plans for purchasing Wildfire enterprise and look at MS LCS again.
Guys - we have updated to Spark 2.5.0 and now we experiance quite a serious issue.
One by one Spark is taking-up 50% of CPU (dual-core machines, hence is 100% on one core) on each of the machines we have upgraded. It does critically impact on PCs performance - they become very unstable and slow.
Only solution is to close spark and startup it again, it than works ok for a while untill i thappens again.
This looks very serious, especialy as this suppose to be the STABLE release.
We have not had that issue while using Spark Betas we have been on ealier.
Just to mention we use Spark .exe installer with bundled JRE
I’‘d like to apologize for the slipped feature. We really thought we could get it done, but didn’‘t do enough homework up front to realize just how complicated the implementation was going to be. I’‘m sorry that we didn’'t reset expectations sooner.
As Matt mentioned we are moving to a release train model with Spark. What that means is that we’‘ll have a more predictable release cycle and we’‘ll be turning the estimation process around to see what we can get done in each 3 week cycle rather than estimating out a bunch of features and setting a specific date. Our intention with the release train is to avoid situations like this and know further in advance when we are going to miss our goals. In addition, we’'ll be much more focused for each release since there are only 3 weeks in which to develop.
I don’‘t think we’‘ve seen the CPU spike problem before. Are there any errors in the logs or other clues as to what’'s going on? Are you seeing this on a single machine, or multiple?