Re: [Geotools-devel] Jenkins geotools-master-online-postgis

2018-06-02 Thread Andrea Aime
Hi Ben, thanks a lot for the effort! I'm looking at the label failure and it's... well... weird? I'm guessing it has something to do with turning the label upside/down so that it can be read but... why that happens only on that particular machine I cannot fathom Cheers Andrea On Fri, Jun 1,

Re: [Geotools-devel] Jenkins geotools-master-online-postgis

2018-05-31 Thread Ben Caradoc-Davies
Jenkins geotools-master-online-postgis is now passing. The gt-render failures on apollo-slave-02 are not fixed, and I have not yet determined the cause, despite detailed comparison of the font configurations on apollo-slave-01 (where it passes) and apollo-slave-02 (where it fails). The

Re: [Geotools-devel] Jenkins geotools-master-online-postgis

2018-05-30 Thread Nuno Oliveira
Ah! Thank you Ben for the update, much appreciated. On 05/29/2018 11:40 PM, Ben Caradoc-Davies wrote: Nuno, Jenkins apollo-slave-02, the only node used for geotools-master-online-postgis, had an /etc/hosts entry that did not match its DHCP-assigned private network IP address. Fixing this

Re: [Geotools-devel] Jenkins geotools-master-online-postgis

2018-05-29 Thread Ben Caradoc-Davies
Looks like a severe font problem to me. This test is supposed to be using the Bitstream Vera Sans we include with GeoTools. Attached screenshot made by sshing into apollo-slave-02 with X forwarding and manual cookie copying, after installing xorg-x11-xauth. As jenkins in

Re: [Geotools-devel] Jenkins geotools-master-online-postgis

2018-05-29 Thread Ben Caradoc-Davies
Torben, although the build still fails, your guess was good because "fc-match sans" matches DejaVu on both slaves (LGC just has a smaller code point coverage): apollo-slave-01: DejaVuSans.ttf: "DejaVu Sans" "Book" apollo-slave-02: DejaVuLGCSans.ttf: "DejaVu LGC Sans" "Book" In other

Re: [Geotools-devel] Jenkins geotools-master-online-postgis

2018-05-29 Thread Ben Caradoc-Davies
Thanks, Torben. We include Bitstream Vera fonts and some tests depend on them to make them more robust against font differences. I do not think any of the rendering tests use DejaVu fonts but it should not hurt to try. Kind regards, Ben. On 30/05/18 11:45, Torben Barsballe wrote: I've

Re: [Geotools-devel] Jenkins geotools-master-online-postgis

2018-05-29 Thread Torben Barsballe
I've installed the following font packages on apollo-slave-02, and restarted the geotools-master-online-postgis build; lets see if that works - dejavu-sans-mono-fonts - dejavu-lgc-sans-fonts - dejavu-fonts-common If not, it might be worth it to take a look at what packages

Re: [Geotools-devel] Jenkins geotools-master-online-postgis

2018-05-29 Thread Torben Barsballe
If its label tests failing, it may mean that some fonts aren't installed on the build box - I recall that CentOS tends to have a very different set of default fonts then most other systems (Ubuntu, OS X, etc). Let me see if I can find a useful list of fonts to install... Torben On Tue, May 29,

[Geotools-devel] Jenkins geotools-master-online-postgis

2018-05-29 Thread Ben Caradoc-Davies
Nuno, Jenkins apollo-slave-02, the only node used for geotools-master-online-postgis, had an /etc/hosts entry that did not match its DHCP-assigned private network IP address. Fixing this fixed HttpConfigTest in Jenkins geotools-master-online-postgis. This will probably also fix

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-05-02 Thread Ben Caradoc-Davies
Yes, I think Daniele made the hstore tests robust against the absence of the hstore extension. Kind regards, Ben. On 02/05/18 11:09, Torben Barsballe wrote: As of the latest tests, it seems like the H2 failure is the only thing causing the geotools-master-online-postgis tests to fail - the

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-05-02 Thread Torben Barsballe
As of the latest tests, it seems like the H2 failure is the only thing causing the geotools-master-online-postgis tests to fail - the PostGIS tests pass fine. (As an aside, I was able to "create extension hstore" on the PostGIS DB, but the PostGIS tests were passing before that) Torben On Wed,

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-04-18 Thread Torben Barsballe
On Wed, Apr 18, 2018 at 7:01 AM, Andrea Aime wrote: > Hum: > > Tests run: 4, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 12.32 sec <<< > FAILURE! - in org.geotools.data.h2.H2DataStoreFactoryTest > testTCP(org.geotools.data.h2.H2DataStoreFactoryTest) Time

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-04-18 Thread Andrea Aime
Hum: Tests run: 4, Failures: 0, Errors: 1, Skipped: 0, Time elapsed: 12.32 sec <<< FAILURE! - in org.geotools.data.h2.H2DataStoreFactoryTest testTCP(org.geotools.data.h2.H2DataStoreFactoryTest) Time elapsed: 12.316 sec <<< ERROR! org.h2.jdbc.JdbcSQLException: Exception opening port H2 TCP

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-04-18 Thread Daniele Romagnoli
Hi again, Good news: The 2 test classes (HSTORE and JSON) aren't failing anymore: https://build.geoserver.org/view/geotools/job/geotools-master-online-postgis/571/console However that the build is still failing on different H2 related stuff which seems failing since long time too. Cheers,

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-04-18 Thread Daniele Romagnoli
Hi everybody. >From my previous understanding, the https://osgeo-org. atlassian.net/browse/GEOT-5876 JIRA was closed as "Not a bug" so I didn't spend additional time on it. However, it looks like is still causing issues with online tests so I was able to replicate it locally and made a couple of

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-04-17 Thread Ben Caradoc-Davies
All I know is that I needed a one-time "create extension hstore;" to turn on hstore support in my test database. I do not know anything about the JSON failure. Kind regards, Ben. On 18/04/18 07:23, Torben Barsballe wrote: As far as I can tell, this is a GeoTools failure - the PostGIS

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-04-17 Thread Andrea Aime
Hi Torben, I'm cc'ing Daniele that implemented those tests. As far as I know one has to install two extensions in postgresql in order for those tests to pass, the hstore, and the json data types in particular, which are plugins. I thought the tests were supposed to check if the extensions are

Re: [Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-04-17 Thread Torben Barsballe
As far as I can tell, this is a GeoTools failure - the PostGIS instance does seem to be working (I know this has been an issue in the past, but I don't think PostGIS is the problem this time) I'm not familiar enough with the failing tests (PostGISHStoreOnlineTest and PostGISJsonOnlineTest) to say

[Geotools-devel] Jenkins geotools-master-online-postgis broken for six months

2018-03-30 Thread Ben Caradoc-Davies
https://build.geoserver.org/view/geotools/job/geotools-master-online-postgis/ Builds that are already broken do cause new breakage to be noticed: https://osgeo-org.atlassian.net/browse/GEOT-5981 https://github.com/geotools/geotools/pull/1846 -- Ben Caradoc-Davies Director

Re: [Geotools-devel] Jenkins geotools-master-online-postgis failing for two months

2017-11-25 Thread Andrea Aime
Hi, confused here, looking at the last build the test failures are in H2? Tests in error: H2DataStoreFactoryTest.testTCP:94 » JdbcSQL Exception opening port H2 TCP Serv... H2JNDIDataSourceTest>OnlineTestCase.run:123->JDBCJNDIDataSourceOnlineTest.testJNDIDataSource:55 » SQLNested

Re: [Geotools-devel] Jenkins geotools-master-online-postgis failing for two months

2017-11-24 Thread Torben Barsballe
Looks like the previous failures were in gt-render: Tests run: 3, Failures: 1, Errors: 0, Skipped: 0, Time elapsed: 0.421 sec <<< FAILURE! - in org.geotools.renderer.lite.LabelSpacingTest testLabelsIncreaseWordSpacing(org.geotools.renderer.lite.LabelSpacingTest) Time elapsed: 0.144 sec <<<

[Geotools-devel] Jenkins geotools-master-online-postgis failing for two months

2017-11-23 Thread Ben Caradoc-Davies
Jenkins geotools-master-online-postgis has been failing for two months and is now failing to report real breakage: https://build.geoserver.org/view/geotools/job/geotools-master-online-postgis/ Real breakage: [GEOT-5888] PostGISJsonOnlineTest.testEntryWithSpaces fails

Re: [Geotools-devel] Jenkins geotools-master-online-postgis failures

2017-08-02 Thread Ben Caradoc-Davies
Thanks, Torben. That should do the trick. Kind regards, Ben. On 03/08/17 05:24, Torben Barsballe wrote: Looks like we are; all the jobs run @daily. I've updated the configuration so the two servers don't overlap the online builds ("H H(0-5) * * *" for ares, "H H(12-17) * * *" for

Re: [Geotools-devel] Jenkins geotools-master-online-postgis failures

2017-08-02 Thread Torben Barsballe
Looks like we are; all the jobs run @daily. I've updated the configuration so the two servers don't overlap the online builds ("H H(0-5) * * *" for ares, "H H(12-17) * * *" for build.geoserver.org) Torben On Tue, Aug 1, 2017 at 6:48 PM, Ben Caradoc-Davies wrote: > I have

[Geotools-devel] Jenkins geotools-master-online-postgis failures

2017-08-01 Thread Ben Caradoc-Davies
I have seen a couple of strange failures of geotools-master-online-postgis. Are we running this job on both ares and the new build server at the same time with the same test database? I think these builds are interfering with each other. Kind regards, -- Ben Caradoc-Davies