jsoref spell checker: remove resource from validation

[romani]

Folders to skip by spell checker:
src/test/resources/*
src/it/resources/*
such folders contains ugly code by design, all mistakes in them are done on purpose for testing.

cleanup https://github.com/checkstyle/contribution/blob/master/jsoref-spellchecker/whitelist.words from odd names.

[romani]

@jsoref , can you help us with this issue ?

[jsoref]

Sure, offhand the place to stick this is near the find argument, iirc it currently excludes "images", you could use the same style for "resources" assuming you're willing to exclude all folders named "resources".

I'll look in more detail tomorrow.

[romani]

yes, you might already have tool/command that can easily remove all not found words from whilelist.
there are many resources, some of them are good to be validated by spellchecker, we will skip only src/test/resources/*, src/it/resources/*.

[romani]

@jsoref , thanks a lot for filter .
can you send us PR to cleanup whiltelist ?

I want to activate your tool on PR validation. Right now it is disabled(always return 0 as exit code)

[jsoref]

@romani sure: checkstyle/contribution#296 -- it's pretty small :-)

[jsoref]

If you're asking about for this PR, that was already done in #5693 (reference)

[romani]

The biggest problem of whitelist is that it too much words from our inputs/resources, example https://github.com/checkstyle/contribution/blob/master/jsoref-spellchecker/whitelist.words#L3585

✔ ~/java/github/checkstyle/checkstyle [master|✔] 
$ grep -Rl "ZZZZZZ" *
src/test/resources/com/puppycrawl/tools/checkstyle/checks/javadoc/abstractjavadoc/InputAbstractJavadocPositionWithSinglelineComments.java

as we get rid of them, it should become small it might become reasonable to host it main repo.
ZZZZZZ and similar , can not be placed in whitelist.

[romani]

@jsoref , whitelist file still contains odd words, please help to cleanup.
I did several changes to whitelist to resolve diffs at CIs.

[jsoref]

Hmm...
We could drop:

• single letter repetitions (e.g. ZZZZZ)
• unicode things: u[a-f]+
• hex things: [a-f]+ (I'm not terribly comfortable with this one, it could easily result in letting a misspelling through)

I think this set should be a pretty good win...

[romani]

@jsoref , one more nuance:

~/java/github/checkstyle/contribution/jsoref-spellchecker [master L|✔] $ grep -C 2 "fffff" whitelist.words 
xfffb
xffff
xffffffff
xffffffffffffffff

~/java/github/checkstyle/checkstyle [master L|✔] $ grep -C 2 -r "xffffffffffffffff" *
.../MagicNumberCheckTest.java-            "103:36: " + getCheckMessage(MSG_KEY, "-2.5"),
.../MagicNumberCheckTest.java-            "109:34: " + getCheckMessage(MSG_KEY, "0xffffffff"),
.../MagicNumberCheckTest.java:            "110:36: " + getCheckMessage(MSG_KEY, "0xffffffffffffffffL"),

and works are validated in URLs (should be skipped)

~/java/github/checkstyle/contribution/jsoref-spellchecker [master L|✔] $ grep -C 2 "CHDCHAHD" whitelist.words 
CHDCEAHH
CHDCFJDG
CHDCHAHD
CHDCHBAE
CHDDCDHH

~/java/github/checkstyle/checkstyle [master L|✔] $ grep -r "CHDCHAHD" *
src/main/java/com/puppycrawl/tools/checkstyle/api/JavadocTokenTypes.java:     * <a href="https://docs.oracle.com/javase/8/docs/technotes/tools/unix/javadoc.html#CHDCHAHD">

How can we skip hex stuff from validation ?

[romani]

one more prove that resources are not excluded

$ grep -rl "soooooooooooooooooooooooooooooooooooolongfooooooooooooooooooooooooooooooooooooooooooo" *
src/it/resources/com/google/checkstyle/test/chapter3filestructure/rule32packagestate/InputLineLength.java
src/it/resources/com/google/checkstyle/test/chapter3filestructure/toolongpackagetotestcoveragegooglesjavastylerule/InputLineLength.java

[jsoref]

Ok, the problem is that the files don't start with a leading / and thus the exclude isn't working...

[romani]

issue is resolved.