#4149 - FindBugs fixes #3
Conversation
Noticed by FindBugs. See ome#4149.
Noticed by FindBugs. See ome#4149.
Noticed by FindBugs. See ome#4149.
This allows it to find absolutely all of the classes that it needs. Usually, it is just the slf4j/log4j JARs that are needed (and not included in the compile-time classpath).
No component generates loci-io-image.jar, so it shouldn't be on the classpath. This prevents FindBugs from failing because it can't locate loci-io-image.jar.
Politely requested by FindBugs.
Noticed by FindBugs.
Noticed by FindBugs.
Noticed by FindBugs.
Noticed by FindBugs.
Noticed by FindBugs.
...since we also overrode Object.equals(Object). Noticed by FindBugs.
|
The only thing I can see is some remaining, non-constant UTF-8s. @chris-allan, did you get notified of this PR? |
...so now we are not writing "UTF-8" as a non-constant in many different places.
|
Non-constant usages of "UTF-8" should be resolved now. The default encoding is now defined once in a new class (loci.common.Constants) and used across all of the components. |
|
@joshmoore: Sure did. For what it's worth explicitly setting a "UTF-8" encoding on any of these transforms is bound to potentially introduce or fix several things. It is not the default encoding on Mac OS X for instance. |
|
Fair enough. UTF-8 is the default encoding on Linux (at least for me), though, so I'd be a little surprised if it causes major problems. If you would rather not set encoding to UTF-8 across the board, then we'll need to convince FindBugs not to complain when the default encoding is used. |
|
Doh. Waited so long that this no automatically merges. Might have to take care of manually. Was there any decision on the UTF-8 front? |
|
@joshmoore @chris-allan and I talked about it last week, and the decision was that I need to write some tests to compare UTF-8, MacRoman, and Windows-1252 encodings (the default encodings on various platforms) before we go further. If it's easier, I can close this PR, fix up the branch so that it can be automatically merged, and then re-open once tests are done and pushed. |
|
Hold off on the closing/re-opening, Melissa. Chris suggested a better way to take care of those that we can try out when you are ready. |
These tests will fail on all of the test strings that contain non-ASCII characters.
|
Encoding tests added in above commit. Enforcing UTF-8 across the board will mean that any files encoded with MacRoman or Windows-1252 will be decoded incorrectly if any non-ASCII characters are present. But then, we would have had this problem anyway when reading a MacRoman file on anything other than Mac OS 9 or when reading a Windows-1252 file on anything other than Windows. |
|
@melissalinkert: This one also needs the same |
Conflicts: components/bio-formats/src/loci/formats/in/NativeND2Reader.java
Fix test suite's logback.xml to write to a logfile
Remaining fixes: For small wavelengths and light source linking
Add intermediate ReaderTest class to initialize reader and default sizes
Different transform sequence
Remove unnecessary image count check
Fix formatting for list, note & menu options
Add test to check that explicitly disabling tiling works
KLB reader
Resolves all bugs/warnings found by FindBugs, and updates the 'findbugs' Ant target to fail if a bug or warning is found.
Should close #4149.