[ https://issues.jboss.org/browse/JBIDE-19641?page=com.atlassian.jira.plugi... ] Brian Stansberry commented on JBIDE-19641: ------------------------------------------ [~kabirkhan] I'm not sure what problem your recent comments are addressing. My impression was the WFLY-669 fix (https://github.com/emuckenhuber/wildfly/commit/03b0c954143bb61064c6095c6d...) avoids the issue with InputStreamEntry.InMemoryEntry, at least for the call paths JBDS is using. Oh, I see. Your test wasn't using the call path JBDS does. Yes, if you pass in an InputStream to the DeploymentPlan it needs to read the whole thing to determine the size. We should improve that but AFAIK that doesn't impact JBDS. The chunking algorithm I used for sending response streams to the client is smarter. If we can use that while remaining compatible with older servers that would be better. Another option is the unused InputStreamEntry.CachingStreamEntry. Copying to disk makes me vaguely uncomfortable though. -- This message was sent by Atlassian JIRA (v6.3.15#6346)