[
https://issues.apache.org/jira/browse/CODEC-55?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Will Pugh updated CODEC-55:
---------------------------
>Findbugs points out:
>* the field RefinedSoundex.US_ENGLISH_MAPPING needs to be private otherwise it
>can be changed outside the class.
>* the constructor parameter mapping is externally mutable. The data should be
>copied to private array.
>
>There are similar problems with the Soundex class.
I noticed these from looking at the code, and don't think we should fix them.
Reasons are:
1) People might have existing code using
RefinedSoundex.US_ENGLISH_MAPPING, changing scope could break them. Also, this
is a crazy thing to expect someone to do. Someone changing this mapping is
probably going to break single threaded code as well as multi-threaded code.
2) Right now creating new classes is super quick (probably less than 20
instructions), adding an array copy in there would slow things down
considerably to protect against a case that I don't think is very likely, and
definitely a bad idea in either a single or muliti-threaded environment.
In addition, I think I should revert my change that made URLCodec's charset and
ESCAPE_CHAR members final. Existing code may be depending on this, and these
changes don't help much. They don't help because:
1) ESCAPE_CHAR does not seem to be actually used by URLCodec
2) When calling createThreadSafeCodec you are getting an instance, so we
don't have to worry about the instance being sub-classed and having that
subclass change charset. In the case that someone already has a subclass of
URLCodec, they should be allowed to, but they need to be responsible for making
sure it's thread safe.
> make all "business" method implementations of public API thread safe
> ---------------------------------------------------------------------
>
> Key: CODEC-55
> URL: https://issues.apache.org/jira/browse/CODEC-55
> Project: Commons Codec
> Issue Type: Wish
> Reporter: Qingtian Wang
> Attachments: concurrentCodecs.diff, concurrentQDiff.diff,
> urlcodec.patch
>
>
> Maybe most of the implementations are already thread safe. Just such that
> codec can say so in general...
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
