[
https://issues.apache.org/jira/browse/OAK-4857?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15527901#comment-15527901
]
Alexander Klimetschek edited comment on OAK-4857 at 9/28/16 12:49 AM:
----------------------------------------------------------------------
Basically:
* the entire [space separator "Zs"
category|http://www.fileformat.info/info/unicode/category/Zs/list.htm], with
the exception of the "normal" {{u20}} space, is not allowed in the middle, but
at start or end of a node name (these are {{Character.isSpaceChar}} in Java)
* for regular {{u20}} spaces this is reversed, these are not allowed at the
beginning or end, while allowed in the middle (there is an [extra
check|https://github.com/apache/jackrabbit-oak/blob/2f85bd78b53851fd00d49695712da3094baeb59e/oak-core/src/main/java/org/apache/jackrabbit/oak/plugins/name/Namespaces.java#L260]
just for this space)
* there is a third category, non breaking spaces ({{ua0}}, {{u2007}} and
{{u202f}}), which are allowed everywhere (these are {{isSpaceChar}} but not
{{isWhitespace}})
"True" whitespace such as tabs or newlines are not allowed anywhere, since
OAK-3412.
Note the {{oak.allowOtherWhitespaceChars}} setting introduced in OAK-3412 does
not make a difference, setting it to true gives the pre-1.4 behavior, which
actually allowed _whitespace_ such as newlines everywhere, while it still
prevents all Zs spaces.
For reference, Jackrabbit 2
[seems|https://github.com/apache/jackrabbit/blob/trunk/jackrabbit-spi-commons/src/main/java/org/apache/jackrabbit/spi/commons/conversion/PathParser.java#L257]
to have the same behavior as Oak post OAK-3412.
See also [this similar
comment|https://issues.apache.org/jira/browse/OAK-3412?focusedCommentId=14991336&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14991336]
by [~mevinay].
Technically, things can be slightly confusing in Java with the meaning of
[Character.isSpaceChar()|https://docs.oracle.com/javase/7/docs/api/java/lang/Character.html#isSpaceChar(char)]
vs.
[Character.isWhitespace()|https://docs.oracle.com/javase/7/docs/api/java/lang/Character.html#isWhitespace(char)].
The latter includes the former for the most part, but with the 3 non breaking
spaces as an exception, and then adds all the newline etc. whitespace chars. I
would be arguing for treating all characters of {{Character.isSpaceChar()}}
like the normal {{u20}} space.
was (Author: alexander.klimetschek):
To be exact, the entire [space separator "Zs"
category|http://www.fileformat.info/info/unicode/category/Zs/list.htm], with
the exception of the "normal" {{u20}} space, is affected. For regular {{u20}}
spaces this is reversed, these are not allowed at the beginning or end, while
allowed in the middle. Whitespace such as tabs or newlines are not allowed
anywhere, since OAK-3412.
Note the {{oak.allowOtherWhitespaceChars}} setting introduced in OAK-3412 does
not make a difference, setting it to true gives the pre-1.4 behavior, which
actually allowed _whitespace_ such as newlines everywhere, while it still
prevents all Zs spaces.
For reference, Jackrabbit 2
[seems|https://github.com/apache/jackrabbit/blob/trunk/jackrabbit-spi-commons/src/main/java/org/apache/jackrabbit/spi/commons/conversion/PathParser.java#L257]
to have the same behavior as Oak post OAK-3412.
See also [this similar
comment|https://issues.apache.org/jira/browse/OAK-3412?focusedCommentId=14991336&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-14991336]
by [~mevinay].
Technically, things can be slightly confusing in Java with the meaning of
[Character.isSpaceChar()|https://docs.oracle.com/javase/7/docs/api/java/lang/Character.html#isSpaceChar(char)]
vs.
[Character.isWhitespace()|https://docs.oracle.com/javase/7/docs/api/java/lang/Character.html#isWhitespace(char)].
The latter includes the former for the most part, but with a few exceptions,
and then adds all the newline etc. whitespace chars. I would be arguing for
treating all characters of {{Character.isSpaceChar()}} like the normal {{u20}}
space.
> Support space chars common in CJK inside node names
> ---------------------------------------------------
>
> Key: OAK-4857
> URL: https://issues.apache.org/jira/browse/OAK-4857
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: core
> Affects Versions: 1.4.7, 1.5.10
> Reporter: Alexander Klimetschek
> Attachments: OAK-4857-tests.patch
>
>
> Oak does not allow spaces commonly used in CJK like {{u3000}} (ideographic
> space) or {{u00A0}} (no-break space) _inside_ a node name, while allowing
> them at the _beginning or end_.
> They should be supported for better globalization readiness, and filesystems
> allow them, making common filesystem to JCR mappings unnecessarily hard.
> Escaping would be an option for applications, but there is currently no
> utility method for it
> ([Text.escapeIllegalJcrChars|https://jackrabbit.apache.org/api/2.8/org/apache/jackrabbit/util/Text.html#escapeIllegalJcrChars(java.lang.String)]
> will not escape these spaces), nor is it documented for applications how to
> do so.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
