[
https://issues.apache.org/jira/browse/HTTPCLIENT-427?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12865134#action_12865134
]
Sam Berlin commented on HTTPCLIENT-427:
---------------------------------------
One statement, one question:
1) Without looking at the code involved, the AtomicLong variable would only
need to be volatile if the variable itself were being assigned to a new
AtomicLong in one thread and used in another. If the AtomicLong object is
created & assigned in one place (preferably the constructor, allowing the
object to be final), then calling 'set' or 'get' on it is effectively the same
as calling get or set on a volatile long. If you say 'new AtomicLong'
somewhere other than the constructor, something is probably wrong with the code.
2) Re: the byte[].clone() that sebb added. Again I haven't looked too deeply
at the code (just the diffs), but is that the byte[] of the body of a cached
response? If so, that could potentially be very large. Cloning it could lead
to memory crunches. Is it possible to change the objects involved to
ByteBuffers & convert it using asReadOnly? Additionally, if it is a byte[] of
the body of cached response... I'm concerned that larger responses (ie,
downloading a large file) wouldn't even be able to be stored in memory. Does
the API allow for reading the data from an InputStream (or Channel?). (If that
byte[] is not for the body of cached responses, you can ignore this question.)
> Implement a cache to perform real request only when needed
> ----------------------------------------------------------
>
> Key: HTTPCLIENT-427
> URL: https://issues.apache.org/jira/browse/HTTPCLIENT-427
> Project: HttpComponents HttpClient
> Issue Type: Improvement
> Components: HttpClient
> Environment: Operating System: Linux
> Platform: PC
> Reporter: Marc Guillemot
> Fix For: 4.1 Alpha2
>
> Attachments: caching_client.zip,
> HTTPCLIENT-427-Immutable-CacheEntry2.patch,
> HTTPCLIENT-427-immutable-cachenentry.patch, httpclient-cache.zip
>
>
> Browsers may cache received content according to the values of different
> response headers. It would be great if HttpClient could do the same.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
