Uploaded image for project: 'Lucene - Core'
  1. Lucene - Core
  2. LUCENE-1852

Fix remaining localization test failures in lucene

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 2.9
    • Component/s: None
    • Labels:
      None
    • Lucene Fields:
      New

      Description

      see also LUCENE-1836 and LUCENE-1846

      all tests should pass under different locales.
      the fix is to run 'ant test' under different locales, look and fix problems, and use the LocalizedTestCase from LUCENE-1836 to keep them from coming back.

      the same approach as LUCENE-1836 fixes the core queryparser, but I am running ant test under a few locales to look for more problems.

        Issue Links

          Activity

          Hide
          rcmuir Robert Muir added a comment -

          patch needs the LocalizedTestCase impl from LUCENE-1836

          Show
          rcmuir Robert Muir added a comment - patch needs the LocalizedTestCase impl from LUCENE-1836
          Hide
          rcmuir Robert Muir added a comment -

          all remaining locale-dependent test failures in core/contrib to my knowledge
          only tested a few locales as it takes some time.

          (but when issues were found, converted to LocalizedTestCase which tests all locales)

          Show
          rcmuir Robert Muir added a comment - all remaining locale-dependent test failures in core/contrib to my knowledge only tested a few locales as it takes some time. (but when issues were found, converted to LocalizedTestCase which tests all locales)
          Hide
          lafa Luis Alves added a comment -

          I reviewed the code, all changes look good to me.

          I tested it and it looks good .

          Michael can you commit this patch with LUCENE-1836, both are related and this issue prereqs it.

          Show
          lafa Luis Alves added a comment - I reviewed the code, all changes look good to me. I tested it and it looks good . Michael can you commit this patch with LUCENE-1836 , both are related and this issue prereqs it.
          Hide
          rcmuir Robert Muir added a comment -

          Luis, thanks for the review. yes this one needs LUCENE-1836 committed first, because it has LocalizedTestCase.

          Show
          rcmuir Robert Muir added a comment - Luis, thanks for the review. yes this one needs LUCENE-1836 committed first, because it has LocalizedTestCase.
          Hide
          thetaphi Uwe Schindler added a comment -

          Looks good.
          Robert, did you find any other failing tests? If not, we could commit.

          Show
          thetaphi Uwe Schindler added a comment - Looks good. Robert, did you find any other failing tests? If not, we could commit.
          Hide
          rcmuir Robert Muir added a comment -

          Uwe, no.

          But i only tested under the following:

          • th_TH (Thai)
          • ko_KR (Korean)
          • ar_JO (Arabic)

          but any problems under these (only found issues under thai and korean) and I switched to LocalizedTestCase so those components test all locales.

          Show
          rcmuir Robert Muir added a comment - Uwe, no. But i only tested under the following: th_TH (Thai) ko_KR (Korean) ar_JO (Arabic) but any problems under these (only found issues under thai and korean) and I switched to LocalizedTestCase so those components test all locales.
          Hide
          michaelbusch Michael Busch added a comment -

          Shall I commit this one too?

          Show
          michaelbusch Michael Busch added a comment - Shall I commit this one too?
          Hide
          rcmuir Robert Muir added a comment -

          in my opinion the issue should be resolved.

          I tested the problematic locales (different calendars, wacky date formats with whitespace), and the changes are identical to LUCENE-1836 (only different queryparsers).
          for each of these queryparsers now all locales are run under the tests, so we should be good to go.

          (But i honestly cannot say i ran 'ant test' under all 50 locales!)

          Show
          rcmuir Robert Muir added a comment - in my opinion the issue should be resolved. I tested the problematic locales (different calendars, wacky date formats with whitespace), and the changes are identical to LUCENE-1836 (only different queryparsers). for each of these queryparsers now all locales are run under the tests, so we should be good to go. (But i honestly cannot say i ran 'ant test' under all 50 locales!)
          Hide
          michaelbusch Michael Busch added a comment -

          great work and testing!

          I'll commit shortly...

          Show
          michaelbusch Michael Busch added a comment - great work and testing! I'll commit shortly...
          Hide
          rcmuir Robert Muir added a comment -

          i will mention under some of these locales, back compat tests will fail for the same reasons.

          it is my understanding these are the tests from 2.4.x against the current code?, don't think its wise to fix them

          in the next release the back compat tests will pass if we fix them now, right?

          Show
          rcmuir Robert Muir added a comment - i will mention under some of these locales, back compat tests will fail for the same reasons. it is my understanding these are the tests from 2.4.x against the current code?, don't think its wise to fix them in the next release the back compat tests will pass if we fix them now, right?
          Hide
          michaelbusch Michael Busch added a comment -

          Committed revision 807499.

          Show
          michaelbusch Michael Busch added a comment - Committed revision 807499.
          Hide
          michaelbusch Michael Busch added a comment -

          it is my understanding these are the tests from 2.4.x against the current code?, don't think its wise to fix them

          in the next release the back compat tests will pass if we fix them now, right?

          Yeah, IMO we don't need to fix them anymore. We're close to the release, so soonish the ones you just fixed will be the back compat ones.

          Show
          michaelbusch Michael Busch added a comment - it is my understanding these are the tests from 2.4.x against the current code?, don't think its wise to fix them in the next release the back compat tests will pass if we fix them now, right? Yeah, IMO we don't need to fix them anymore. We're close to the release, so soonish the ones you just fixed will be the back compat ones.

            People

            • Assignee:
              michaelbusch Michael Busch
              Reporter:
              rcmuir Robert Muir
            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development