[ANN] Agra EAP, build 3914, is available

Hello everyone,

We've prepared another almost bugfix-only EAP build.

- Delayed commit and test view in Eclipse plugin.
Those of you who work in teams with Eclipse users - please give it a try!
- Improved Web Diff
- True AJAX refresh of dashboard/overview page
- Improved (once more) My Changes view
- Many bugfixes and improvments

Please see screenshots of the new stuff at
http://www.jetbrains.net/confluence/display/TW/ReleaseNotesforAgra%2Cbuild+3914

Please read the full list of changes at: http://tinyurl.com/sk9y6

And please download (distribution size became smaller once more):
http://www.jetbrains.net/confluence/display/TW/Download+Latest

Important Note: we recommend using Mysql in production environment with TeamCity,
because we've had several reports when HSQLDB database crashed under the load.

To setup Mysql in Agra read the following instructions:
http://www.jetbrains.net/confluence/display/TCD/Installation#Installation-SettingupMySQL

We are working hard to improve quality of TeamCity before the beta,
so please let us know which things are most annoying for you and
which issues you really want to get fixed in the Agra release.

With kind regards,
KIR

--
Kirill Maximov
JetBrains, Inc.
http://www.jetbrains.com
"Develop with pleasure!"

14 comments
Comment actions Permalink

Kirill,

I upgraded from 3776 and a failure in loading the native colorer prevents the server to start up even though it is disable in the spring context file as per your instructions (I did not change anything there).

I am running teamcity on a AMD x64 Centos ( 2.6.9-42.ELsmp #1 SMP Tue Aug 15 10:35:26 BST 2006 x86_64 x86_64 x86_64 GNU/Linux). You do not include any compatible binaries so I had to revert to brute force:
I tried to compile from source the colorer on my platform. Unfortunately, after replacing the net_df_colorer.so with my own, I get the following link error
java.lang.UnsatisfiedLinkError: init
at net.sf.colorer.ParserFactory.init(Native Method)
at net.sf.colorer.ParserFactory.(ParserFactory.java:20) at jetbrains.buildServer.web.util.ColorerSyntaxHighlighter.]]>(ColorerSyntaxHighlighter.java:35)

What package should I use? How should I build it? What are these .jnilib files. Should I update them as well? (I used this http://downloads.sourceforge.net/colorer/Colorer-take5-linux.be5.tar.bz2?use_mirror=easynews)

Thanks

Jacques

Message was edited by:
Jacques Morel

0
Comment actions Permalink

I've just got stung by this as well, running on Solaris. (Maybe even a little stranger - it appears to be trying to load the linux .so file and getting really confused by it)

Given that it is providing optional behavior, I hope it'll be possible to run TC if loading the native library fails.

Better yet, how about a pure Java syntax colorer, based in the IDEA one ;)

A workaround so I can at least start this build would be really helpful.

-j

0
Comment actions Permalink

Johannes Graham wrote:

I've just got stung by this as well, running on Solaris. (Maybe even a little stranger - it appears to be trying to load the linux .so file and getting really confused by it)

Given that it is providing optional behavior, I hope it'll be possible to run TC if loading the native library fails.

Better yet, how about a pure Java syntax colorer, based in the IDEA one ;)

A workaround so I can at least start this build would be really helpful.

-j


Everybody who faced native library incompatibility, please remove section

]]>

around line 25 of WEB-INF\buildServerSpring.xml to quickly get rid of an
exception at server startup.

Also you can try to get Colorer library sources and recompile for your
platform. Binaries should go to WEB-INF\lib folder.

As for IDEA based highlighter - it's really not an easy task to strip it
out of IDE and even after it will require support for whole bunch of
.Net, shell and other file types to be written from scratch. Anyway this
may happen at some point in future.

Also if you aware or standalone pure java library with support for great
number of languages/filetypes - give us a notice, we can give it a
chance, if license will be suitable.

--
Alexey Gopachenko
JetBrains Inc.
http://www.intellij.com
"Develop with pleasure!"

0
Comment actions Permalink

Alexey,

Thanks for the workaround. As I have already tried the recompilation from source with no success, could you tell us exactly the source package you used (give us the link to the download or attach the source bundle to the wiki)?

Thanks

0
Comment actions Permalink

Alexey,

Thanks for the workaround. I would still like to make the colorer work. As I have tried to recompile from source without success (see my earlier post), could you tell us what exact source package you used (give us the link to the source bundle or attach the sources bundle to the wiki maybe)?
Also does the .jnilib files need updating? What are they for?

Thanks

0
Comment actions Permalink

Kirill Maximov (JetBrains) wrote:


We are working hard to improve quality of TeamCity before the beta,
so please let us know which things are most annoying for you and
which issues you really want to get fixed in the Agra release.


For me the most important things are:
1) The inspections runner must be running at about the same performance
level as 1.2, at the moment it's catastrophic (when I last looked, I had
a 12x slowdown) http://jetbrains.net/jira/browse/IDEADEV-14356
2) Notifications must be working properly:
http://www.jetbrains.net/jira/browse/TW-1649
http://www.jetbrains.net/jira/browse/TW-1803
http://www.jetbrains.net/jira/browse/TW-1863
3) Make it very clean which JDK it being used at various points in the
process - between "system default" (i.e. found in the path), project
default (specified in ipr file) and agent properties it's not at all
clear and it's only by fiddling that I've got it working.

The notifications are the biggest usability problem for me - I get
enough mail during the day without TeamCity sending me lots more
unnecessary junk. Don't tell me I have caused a test to fail if the test
was already failing before my commit; don't tell me that a test is still
failing if I haven't specifically asked for it.

Of course, it could be that a lot of these are already addressed, I
haven't really played with 3914 yet.

Thanks,
R

0
Comment actions Permalink

I can't download the Eclipse plugin from the "My Settings & Tools" page:
The link says "update site link" instead of "download" (which it says for the other plugins) and when I click on it, an error page appears saying "The requested resource (/update/eclipse/) is not available."

Running on XP.

0
Comment actions Permalink

Jacques Morel wrote:

Alexey,

Thanks for the workaround. I would still like to make the colorer work. As I have tried to recompile from source without success (see my earlier post), could you tell us what exact source package you used (give us the link to the source bundle or attach the sources bundle to the wiki maybe)?
Also does the .jnilib files need updating? What are they for?

Thanks


I've used prebuild binaries and classes from net.sf.colorer 0.8 for
eclipse platform. .jnilib is for MacOSX and .dll is for windows :)

--
Alexey Gopachenko
JetBrains Inc.
http://www.intellij.com
"Develop with pleasure!"

0
Comment actions Permalink

Hello Tobias,

Please copy the "update site link" URL and use it as Eclipse update site. e.g. The same process for Subclipse plugin is outlined at http://subclipse.tigris.org/install.html

Please note that current Eclipse plugin version works with Subversion and assumes Subclipse 1.2.x (from update site http://subclipse.tigris.org/update_1.2.x ) is installed in Eclipse.

--
Best regards,

Yegor Yarko
Quality Assurance Engineer
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"

0
Comment actions Permalink

Ok, I misunderstood that. Maybe a small explanation on the page would help other user as well.

Here's another small glitch:

When I take a look at a build result with the IDEA plugin and I click on the "Failed tests" tab, it just says "Loading tests..." forever if there aren't any failed tests. Noticed that with the Maven2 runner at least.

0
Comment actions Permalink

Tobias,

Yes, we will make the "update site link" link to show explanatory text as it indeed can be confusing. Thanks for noting.

Concerning endless "Loading tests...": there is already an issue to address this:
http://www.jetbrains.net/jira/browse/TW-2052 You can watch it if you like.

--
Best regards,

Yegor Yarko
Quality Assurance Engineer
JetBrains, Inc
http://www.jetbrains.com
"Develop with pleasure!"

0
Comment actions Permalink

Kirill Maximov (JetBrains) wrote:


We are working hard to improve quality of

TeamCity before the beta,

so please let us know which things are *most

annoying for you* and

which issues you really want to get fixed in the

Agra release.


For me the most important things are:
1) The inspections runner must be running at about
the same performance
level as 1.2, at the moment it's catastrophic (when I
last looked, I had
a 12x slowdown)


This problem should be resolved in the nearest EAP.

http://jetbrains.net/jira/browse/IDEADEV-14356
2) Notifications must be working properly:
http://www.jetbrains.net/jira/browse/TW-1649
http://www.jetbrains.net/jira/browse/TW-1803
http://www.jetbrains.net/jira/browse/TW-1863

3) Make it very clean which JDK it being used at
various points in the
process - between "system default" (i.e. found in the
path), project
default (specified in ipr file) and agent properties
it's not at all
clear and it's only by fiddling that I've got it
working.


Please watch
http://www.jetbrains.net/jira/browse/TW-2102


The notifications are the biggest usability problem
for me - I get
enough mail during the day without TeamCity sending
me lots more
unnecessary junk. Don't tell me I have caused a test
to fail if the test
was already failing before my commit; don't tell me
that a test is still
failing if I haven't specifically asked for it.


We plan to write documentaion explaining how notifications work in Agra. Hope this will help.

Really thanks for your feedback!

KIR

0
Comment actions Permalink

Hi Kir,
Thanks for your comments

Kirill Maximov (JetBrains) wrote:
>> 1) The inspections runner must be running at about
>> the same performance level as 1.2


This problem should be resolved in the nearest EAP.


Do you mean 3914 or an upcoming build?
Thanks,
R

0
Comment actions Permalink

I mean build 3989, which is already available for download.

0

Please sign in to leave a comment.