This is harmless, but delays the actual fix, and if we're doing a major release anyway, I don't think we need to do this. Still it'd be an improvement as an intermediate step if we can be sure to get to option 1 eventually.Ĭut a point release with a deprecation warning whenever a timeout isn't provided, and start prepping people for one of the other solutions (which one?) ![]() This feels pretty safe, but I think it would drain momentum from option number 1, so I don't favor it. But I think this is the correct long term fix if this were a cleanroom implementation.Ĭut a new minor version with the OS-based Connect timeout (as in sets a default timeout and resolves #3070 #5779) I strongly favor this option, but there's a bigger picture around releases that I lack visibility into, so maybe it's not practical. We have a number of proposals:Ĭut a new major version with the proper fix (reasonable default timeouts) Over in #5779, there is a bit of discussion about how to fix this, but I think the blocker here is that we don't have direction on what kind of fix would be merged. There are a number of discussions about this elsewhere (search closed issues), and we don't want that conversation to muddy this issue too. Please don't use this issue to discuss adding a timeout attribute to requests. If a large default were provided to all requests and all sessions, what negative impact would that have? The only thing I can think of is that some programs will get timeout exceptions where they previously hung, which seems like an improvement to me. I also see that a lot of people want Session objects to have a timeout parameter, and this might be a way to do that as well. (I'm in the process of updating all my code.) It took one of my programs hanging forever for me to realize that the default here isn't really very good for my purposes. The reason I'm thinking about this is because I've used requests for a few years now and until now I didn't realize the importance of providing a timeout. Make every user configure this in every request - bigger API breakage.I know browsers have a default, so that may be a simple place to begin. ![]() I have a feeling I'm about to get a swift education on this topic, but I've been thinking about the pros/cons of changing requests so that somehow there is a timeout value configured for every request.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |