I am still curious... why was this broken?
I know IGB supports synonyms for chromosome names and I think it is supposed to support "Chr1" and "chr1" for the A. thaliana genome. Might this test have been designed to test that the synonym system is working? And the test failing is indicating that the synonym system is NOT working...? I doesn't say that, but this file is pretty void of comments, so we have to make assumptions about what each test is designed to test for.
In an older version (8.2.3) the chromosome names that display in the A. thaliana genome are lower cased, "chr1".
This is odd, given that IGB doesn't store data. So both versions that I have running next to each other (v8.2 and v9.0) should be looking at the same data, because they are looking at the same remote quickload site. But in v8.2 shows "chr1" and v9.0 shows "Chr1".
The "Chr1" is a more accurate reflection of what is in the genome file. So how does v8.2 have "chr1" ?
The user's guide talks about making a chromosomes.txt file with synonyms for chromosome names.
https://wiki.transvar.org/display/igbman/Personal+Synonyms
I think IGB has a built-in chromosomes.txt file.
I think there must have been a change in that file or the code that uses it.
We should test the synonyms functionality.
The current solution is good. It checks that the expected chromosome names are present (based on the names they have in the sequence file). And making the test pass allows all of the subsequent tests to run.
I tested this issue on the master branch (since the description says that the problem relates to developers running from the master branch) and I tested this issue on the
IGBF-1142branch of Jenny's fork. I realize that the fix for this issue must not be incorporated into the master branch yet, because I am experiencing different behavior on the branches when building from Netbeans.On the
IGBF-1142branch, I am able to successfully build IGB from Netbeans without skipping tests. The build is successful on my Windows and Mac. For this reason, I believe this JIRA story to be resolved.On the master branch I am running into something odd: The build is successful on my Windows machine but fails on my Mac. I am trying to look into why this is but do not have any answers yet. I have reached out to Jenny to see why this may be, but I don't think any fix for this issue has been merged into the master branch yet, so I still believe this JIRA story to be resolved although I do think the behavior is odd.
Since the build is successful on the
IGBF-1142branch on both my Windows and Mac, I am marking this issue as resolved. If I gain new information that the fix should have been merged into the master branch already, I will re-open the issue and continue looking into it.