[
https://issues.apache.org/jira/browse/GEODE-9487?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17394112#comment-17394112
]
Blake Bender commented on GEODE-9487:
-------------------------------------
Looks like the thing we really want is %z, not %Z. The lower-case version
"writes offset from UTC in the ISO 8601 format (e.g. -0430), or no characters
if the time zone information is not available," which is probably always going
to produce a valid utf-8 string.
> Timezone is potentially invalid string in log files for non-English locales
> ----------------------------------------------------------------------------
>
> Key: GEODE-9487
> URL: https://issues.apache.org/jira/browse/GEODE-9487
> Project: Geode
> Issue Type: Improvement
> Components: native client
> Reporter: Blake Bender
> Priority: Major
>
> The native client logger was rewritten several months ago to use, among other
> things, `std::put_time` when formatting log strings, specifically using the
> "%Z" formatting for timezone. The documentation for this formatting says
> "writes locale-dependent time zone name or abbreviation, or no characters if
> the time zone information is not available," and indeed it does. We just
> received a log file from a customer machine in APJ region, i.e. locale is set
> to something like Traditional Chinese. These strings are dumped into the log
> file as is, in some unknown MBCS encoding, and contain invalid utf-8 start
> codes, throwing off our Python-based parsing tool.
> We should probably just drop time zone from our logs, or normalize everything
> somehow. In any case, we need to write valid utf-8, just to play well with
> others.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
