[HTTPCLIENT-2262] Clarify infinite timeout vs disabled timeout

• Type: Task
• Status: Resolved
• Priority: Major
• Resolution: Information Provided
• Affects Version/s: 5.2.1
• Component/s: HttpClient (classic)
• Assignee: Unassigned
• Reporter: Nicolas Filotto
• Created: 14/Mar/23 10:58
• Updated: 24/Jun/23 21:40
• Resolved: 16/Mar/23 12:38

Description

I have a question regarding the expected behavior when setting a timeout to 0.

In HttpClient, If I read for example the Javadoc of the connectionRequestTimeout, it tells me that a timeout of 0 is infinite.

In HttpCore, If I read the Javadoc of Timeout#DISABLED, it tells me that the timeout is simply disabled when set to 0 but what disabled means precisely?

If I keep digging in HttpCore, I see several classes like StrictConnPool that leverage Future and Lock for which a timeout set to 0, means that we expect an immediate result either with a success or a TimeoutException, so I assume that it is what "disabled timeout" means but if so it would be in conflict with an infinite timeout as I personally don't expect to get a TimeoutException in case of an infinite timeout but maybe I'm wrong and it just needs to be clarified?