[
https://issues.apache.org/jira/browse/TIKA-4920?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18125662#comment-18125662
]
ASF GitHub Bot commented on TIKA-4920:
--------------------------------------
Copilot commented on code in PR #3315:
URL: https://github.com/apache/tika/pull/3315#discussion_r4233917070
##########
tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-font-module/src/test/java/org/apache/tika/parser/font/FontParsersTest.java:
##########
@@ -72,6 +73,8 @@ public void testAdobeFontMetricParsing() throws Exception {
}
@Test
+ // fontbox TTFDataStream.readInternationalDate builds its Calendar in the
default locale (TIKA-4920)
+ @DisabledIf("org.apache.tika.test.TestLocales#nonGregorianDefault")
Review Comment:
This skips all TTF parsing assertions for non-Gregorian defaults, while only
the CREATED/MODIFIED checks at lines 94–95 depend on FontBox's locale-sensitive
calendar. Content type, names, rights, and extracted text should remain covered
in those nightly runs; condition or split out only the two date assertions.
##########
tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/OutlookParserTest.java:
##########
@@ -200,6 +201,8 @@ public void testOutlookNew() throws Exception {
}
@Test
+ // POI LocaleUtil.getLocaleCalendar dates MAPI times in the default
calendar (TIKA-4920)
+ @DisabledIf("org.apache.tika.test.TestLocales#nonGregorianDefault")
Review Comment:
The locale-sensitive behavior is confined to the five MAPI time assertions
at lines 247–251, but this annotation also suppresses HTML parsing, Chinese
text, recipient, and identifier coverage. Keep those locale-independent
assertions running and condition or split out only the date checks.
##########
tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-pkg-module/src/test/java/org/apache/tika/parser/pkg/RarParserTest.java:
##########
@@ -43,6 +44,8 @@ public class RarParserTest extends AbstractPkgTest {
* fired for all the embedded entries.
*/
@Test
+ // junrar FileHeader.getDateDos uses Calendar.getInstance() (TIKA-4920)
+ @DisabledIf("org.apache.tika.test.TestLocales#nonGregorianDefault")
Review Comment:
This disables the entire embedded-entry traversal test for non-Gregorian
defaults even though only the modified-date assertions at lines 80–83 depend on
junrar's locale-sensitive calendar. As a result, filenames, media types, and
extracted content are no longer covered in those nightly runs. Keep the parse
and locale-independent assertions enabled and condition or split out only the
modified-date checks.
##########
tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-pkg-module/src/test/java/org/apache/tika/parser/pkg/ZipEntryTimesTest.java:
##########
@@ -65,6 +66,8 @@ static void tearDown() {
}
@Test
+ // commons-compress ZipUtil.dosToJavaDate uses Calendar.getInstance()
(TIKA-4920)
+ @DisabledIf("org.apache.tika.test.TestLocales#nonGregorianDefault")
Review Comment:
Only the DOS timestamp assertion at line 92 depends on
`ZipUtil.dosToJavaDate`; the extended-timestamp assertion and metadata-key
checks are locale-independent. Disabling the whole method removes coverage of
those paths under both non-Gregorian nightly builds. Keep the test enabled and
condition or split out only the DOS assertion.
> Figure out why CI didn't catch a locale bug
> -------------------------------------------
>
> Key: TIKA-4920
> URL: https://issues.apache.org/jira/browse/TIKA-4920
> Project: Tika
> Issue Type: Task
> Reporter: Tim Allison
> Priority: Major
>
> [~tilman] found a locale issue failing the build:
> https://issues.apache.org/jira/browse/TIKA-4902?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18118678#comment-18118678
>
> CI should have caught this.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
