Skip to content

Sprint5 server backlog#48

Merged
joshmoore merged 28 commits intoome:developfrom
joshmoore:sprint5-server-backlog
Jan 11, 2012
Merged

Sprint5 server backlog#48
joshmoore merged 28 commits intoome:developfrom
joshmoore:sprint5-server-backlog

Conversation

@joshmoore
Copy link
Copy Markdown
Member

When someone gets the chance, it would be good to start evaluating some of these commits so that before we start the bug work (sprint6) some of this could already be in the mainline.

@jburel
Copy link
Copy Markdown
Member

jburel commented Dec 28, 2011

in order to speed up the review, it will good to indicate for some of the commits what to do e.g the compile one

@jburel
Copy link
Copy Markdown
Member

jburel commented Dec 28, 2011

Build using that branch. Insight now hangs on Starting Caching Service so the problem is not related to DnD.
java -version java version "1.5.0_28" Java(TM) 2 Runtime Environment, Standard Edition (build 1.5.0_28-b04-382-9M3326) Java HotSpot(TM) 64-Bit Server VM (build 1.5.0_28-157, mixed mode)

@jburel jburel mentioned this pull request Dec 28, 2011
@jburel
Copy link
Copy Markdown
Member

jburel commented Dec 29, 2011

Hanging problem also happens on develop.

Using ./build.py -Ddist.dir=... no longer works with the
insight build since the dist.dir property is copied into
the insight build overwriting the expected value. Instead,
set the property in etc/local.properties until the insight
build has been fully refactored.
With other changes to have dist.dir configurable, it became
necessary to merge directories.xml and global.xml so that dist.dir
could be set via local.properties before it was used elsewhere.
This also required changing the OMERODIR to point
at dist.dir so that the Java and SQL files used
in the test could be found.
@joshmoore
Copy link
Copy Markdown
Member Author

To test

bfced25 "Use getResultsString in ServiceHandler"

Check in the server log when using IQuery.projection for nice output like [[1, 2, 3], [4, 5, 6], [7, 8, 9]] rather than [[Ljava.lang.Object;@50dfc362,...

5051327 "Use dir.dir for python test with non-default dist"

./build.py build-dev -Ddist.dir=/tmp/my-build-dir should not touch the omero.git/dist at all. Commit 9627d00 was also necessary to make this work. A typo was fixed in a later commit. Further, the insight build does not permit the use of "-Ddist.dir". Instead, set dist.dir in etc/local.properties. A warning message to this effect is printed:

/Users/cblackburn/Work/repos/ome/components/insight/build/dist.xml:57: ERROR: dist.dir is currently set!

        This most likely occurred because dist.dir was set on the top-level
        build:  ./build.py -Ddist.dir=/tmp/example

        This no longer works with the insight build. Instead, set dist.dir
        in the etc/local.properties file.

3a35d38 "More deletes on 'clean'"

./build.py build-dev; ./build.py clean should leave fewer files lying around.

a9196bc "Adding OmeroPy sql directory for testing"

Run the "test/clitest/db.py" test via: ant test -DTEST=clitest.db.

8d59207 "Disabled test"

Group "broken" wasn't enough to prevent this test from being run. It is now marked "enabled=false".

ba60d28 "Move ice check to top-level and print help"

If you're using ice 3.3 and not set ice.compatibility=3.3 in etc/local.properties, then rather than waiting until slice2java is executed under components/blitz an error message should be shown right away. (Similarly, if 3.3 is set in local.properties and you are using ice3.4)

57fefd4 "Fix java 1.6 compile of omero-shares (gh-46)"

If compiling with Java 5, ./build.py build-dev should work. (Previously broken) This was caused by my having just done "javac *.java" under ome/services/sharing/data. The "make_data.py" which I used to create the omero-shares jars now sets the target explicitly.

63d28eb "Enable IPython 0.11 shell (Fix #7112)"

With a new IPython installed (pip install ipython) bin/omero shell should work.

@joshmoore
Copy link
Copy Markdown
Member Author

The most important test for the new postgresql jar is to use OMERO against an earlier (especially 8.4) postgresql server.

chris-allan and others added 8 commits January 2, 2012 11:37
Setting ice.compatibility is no longer necessary in
etc/local.properties. The default value "auto" will
use the current value of `slice2java --version`. If
a specific value is set, then that will be required.
If omero.version changes between two builds, then
an ivy error is printed:

{{{

    ./build.py
    OMERO_BUILD=b123 ./build.py
    ...
    bad revision found in ivy file (Revision: 4.3.3). Use forcedeliver or
    update.

}}}
Now any build run via ./build.py (or build.bat) will have
the current ice version appended to the omero.version value
(e.g. 4.3.3-ice34)
A typo in the "check_ice" method was hiding the ctx.die
exception. Now, "check_ice" is a top level method:

{{{
    bin/omero admin checkice
}}}

and the value of versions.ice rather than ice.compatibility
is being saved to omero_versions.py
@ximenesuk
Copy link
Copy Markdown
Contributor

All items now tested okay including the various fixes since the testing began.

Now tested against postgresql 8.3.8

joshmoore added a commit that referenced this pull request Jan 11, 2012
@joshmoore joshmoore merged commit 4b3ac95 into ome:develop Jan 11, 2012
This was referenced Jan 11, 2012
snoopycrimecop pushed a commit to snoopycrimecop/openmicroscopy that referenced this pull request Nov 19, 2014
2014-SV3+SV4 Robot and Poodle fixes
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants