[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [tracecompass-dev] Test case failures on Window running maven build

Hi David,

I was able to reproduce a similar problem with Windows short path.

The short path comes from org.junit.rules.TemporaryFolder.

I think we can fix it by applying getCanonicalFile() on the File created in ResourceUtilTest.beforeClass().

There are probably other test cases that would fail in similar fashion. I can make a fix with a global search on TemporaryFolder in our tests, but if you later still notice more issues with your short path, please let us know.

I'm not sure about your last question of missing plug-ins when using -Dmaven.test.skip=true. Maybe someone else can help?


On Wed, Feb 20, 2019 at 5:40 PM David Wootton <dwootton@xxxxxxxxxx> wrote:

I'm back to looking at my problems with incorrect event.interval coloring in some of my traces, where I'm trying to modify one of my data provider classes to handle the coloring. The only way I can think to make this works is to request some changes tothe CallStackDataProvider class, but before I request anything I wanted to try making my changes myself to verify it worked.

I needed to run a Maven build in order to do this, then point my target definition file to that build. When I ran the build as mvn clean install I get error messages like this

ResourceUtilTest.testCopyResource:395->createLinkCopyAndVerify:640->createLinkCopyAndVerify:644->createAndVerifyLink:552->createAndVerifyLink:569->verifyLocation:626 expected:<C:/Users/DAVIDW~1/AppData/Local/Temp/junit8790004266760650741/targetFile> but was:<C:/Users/DavidWootton/AppData/Local/Temp/junit8790004266760650741/targetFile>

It looks like the problem has something to deal with file and directory naming rules on Windows 10 where something seems to be getting confused by Windows long file names and Windows short file names/FAT filenames because I see expected text DAVIDW~1 and resulting DavidWootton

I was doing this with a MingGW shell first but it also fails with a Windows shell. I also tried this with HEAD and 4.2 stable branches.

I tried building using the -Dmaven.test.skip=true flag and the build completed, but when I built my plugin and tried to run it as an Eclipse runtime instance that failed with a bunch of exceptions due to being unable to resolve plugins with *tests as the last part of the plugin nam as well as some non-test plugins like org.eclipse.tracecompass.analysis.graph.ui

Any idea what might be going wrong?


tracecompass-dev mailing list
To change your delivery options, retrieve your password, or unsubscribe from this list, visit