[IMAGING-239] Add inflate (deflate algorithm) to TIFF files

[pauldaustin]

• Adds inflate (deflate algorithm) to TIFF files.
• Also allows passing params to reading directories, so that image data loading can be turned off
• Make getTiffRawImageData public so that the image data can be loaded after the directories.

I was going to do 2 commits but then it got committed by accident.

[coveralls]

Coverage increased (+0.002%) to 74.415% when pulling e655d2b on pauldaustin:feature/deflate into 1e725b0 on apache:master.

[kinow]

From an initial look the code looks good. @pauldaustin could you create a JIRA issue for it, and also include some basic unit tests, please? There's some info on contributing here. Thanks!

[kinow]

Thanks for working on the PR comments so quickly. Could you take a look at the Travis CI broken build? It appears to be an issue with the code formatting.

You can look at the .travis.yml file, and run the same maven targets in your own computer to reproduce it.

[pauldaustin]

Do you know if there are an exported configurations for Eclipse code style, formatting and import ordering. I use that on my own projects so it automatically gets formatted properly when I save.

I avoided using the eclipse auto formatter which would resolve these issues as it re-formatted a bunch of the other code. My code style also sorts methods into a consistent order which if run regularly on a project makes merging changes between branches easier.

[pauldaustin]

Looks like the formatter wouldn't have helped as the spaces were at the end of lines in the javadoc comment and it didn't strip them when I tried the formatter.

[kinow]

Looks like the formatter wouldn't have helped as the spaces were at the end of lines in the javadoc comment and it didn't strip them when I tried the formatter.

👍 I kind of got used to the style used in most Commons projects, so I don't have an imported style/formatter settings for Eclipse (I also use it for Java). Some projects have mixed styles too, mixing old and new, with different rules. Easier to try to copy the existing style and run Maven targets before committing I guess.

New error:

[ERROR] 7002: org.apache.commons.imaging.common.bytesource.ByteSource: Method 'public java.lang.String getFilename()' has been removed

I think an existing method that was public has now been removed? That is not backward compatible - i.e. existing users would have issues unless we released a new major version (next release is still maintenance/alpha).

[garydgregory]

I renamed getFilename() (obvious typo) to getFileName(). Since we do not have a 1.0 release yet, breaking BC is OK. We should get the public names right before 1.0.

[kinow]

I renamed getFilename() (obvious typo) to getFileName(). Since we do not have a 1.0 release yet, breaking BC is OK. We should get the public names right before 1.0.

Is there a way to tell cliir to ignore these errors for now then? So that the PR's checks can all pass

[pauldaustin]

I merged master into my branch to get the other changes.

Maybe this will fix the PR check issues.

[kinow]

@pauldaustin no that won't fix the build. I've created a PR that disables Clirr check for backward compatibility - #45

If/once that gets merged, if you rebase your changes again then the build should pass this time.

Thanks
Bruno

[kinow]

Done @pauldaustin . If you rebase this PR onto master, the tests should finally pass. Sorry for the delay!

[pauldaustin]

I'm giving up, seems to be a bunch of errors I'm not sure how to fix, or are in parts of the code I haven't changed.

[kinow]

@pauldaustin just had a look, and the PR is passing in some JDK's, and the failures appear to be related to Travis-CI. Nothing wrong with your code as far as I can tell, so I believe it's ready to be reviewed and hopefully merged :) 👍 thanks for rebasing it

[pauldaustin]

@kinow That's great news thanks.

[kinow]

Merged, thanks!