Apache HttpComponents

ok2c commented on PR #693: URL: https://github.com/apache/httpcomponents-client/pull/693#issuecomment-3393174100 @arturobernalg This feature was requested by @rschmitt and do not think i need to get involved unless absolutely necessary. I suppose Read more

arturobernalg commented on PR #693: URL: https://github.com/apache/httpcomponents-client/pull/693#issuecomment-3391309446 @rschmitt @ok2c can we merge the minimal path (staggered multihome + config flag) as *experimental* and iterate on the Read more

The javadoc for URLEncodedUtils.encodeFormFields() says ... > Encode/escape www-url-form-encoded content. > Uses the URLENCODER set of characters, rather than the UNRESERVED set; this is for > compatibilty with previous releases, URLEncoder.encode() Read more

Hi olegk —I went with Option 1: a small safeguard that caps the async execution queue and fails fast with RejectedExecutionException once the limit is reached (default remains unlimited). The guard also releases the slot on failed/cancel/releaseResources Read more

arturobernalg opened a new pull request, #728: URL: https://github.com/apache/httpcomponents-client/pull/728 Implements Option 1 from the discussion: add a configurable cap on the `async` execution pipeline to prevent the recursive callback loop Read more

abernal The proposed change-set still looks hack-y to me even without the use of ThreadLocal. I am not even sure it actually fixes the original problem in the first place. Given there is no reproducer it cannot be properly verified. If you want, you Read more

eppdot I think I have an idea what is happening. Basically your code shoves thousands (many thousands?) of requests into the execution pipeline without making any attempt at checking whether or not those requests actually succeed. HttpClient Read more

olegk Thank you for your response. I confirmed that by passing DefaultDisposalCallback to the constructors of StrictConnPool and PoolingAsyncClientConnectionManager, the PoolEntry.discardConnection() method closes the connection with a Read more

The IdleConnectionEvictor is performing a potentially blocking I/O operation while holding a global lock on the pool

iwasaki The defect got fixed [1] in the meantime. The fix will be released in version 5.5.1 coming soon. There is also an experimental feature to do graceful connection disposal from outside the connection pool lock targeted for 5.6 release series. Read more