Rebased and reorganized #223 for Jenkins #238
Conversation
Before this commit, the Lwt_stream API provided no reliable way to detect if a stream had been closed. While on_termination notified the user if the stream was closed in the future, it did not tell the user if the stream had been closed in the past. The is_closed function addresses this problem, allowing the user to detect if a stream has already been closed at any point in time. The closed function returns a thread that will sleep until the stream is closed. Lwt_stream.on_termination is now implemented in terms of Lwt.closed. Lwt_stream.on_termination will be deprecated and ultimately removed in favor of Lwt_stream.closed.
The Lwt_stream module had multiple copies of low-level enqueueing code of two varieties copied in serveral places. This commit moves one variety of enqueing code into a single function called enqueue', and implements the second variety as the function enqueue, in terms of enqueue'.
aantron
changed the title
|
Why were my commits rebased and modified? These commits now contain messages that I did not author, even though I'm listed as the author. Also, why not actually include a merge commit so there's a pointer back to discussion in the commit history? This is also going to complicate merging #239. |
|
Everything in the commit messages is something you authored, only some commits were squashed into one, which has trivial sentence changes to make the sentences flow together after concatenation. The only sentence I added is the overall title, which is a natural consequence of a squash. I think you as author and me as committer reflects this situation accurately. I did this while reviewing the code, because the original order, together with hidden comments in #223, made it rather difficult to clearly see what was going on, especially given that I was not yet as familiar with the surrounding code. Given that I had already done this squash for review, I decided to post and then merge it, without asking you to go through the exercise, in potentially multiple iterations with me, of cleaning up the messy history of #223 to achieve basically the same end result. While I now realize I should have clearly told you my intentions in a comment, there also was a link to this PR from #223 that was readily visible. Perhaps the title "...for Jenkins" was misleading? Anyway, I assumed that there had been time to object, so I apologize. There are links back to the discussion. For example, e4f39b8 links to this PR, which links to #223. I grant that this isn't so useful for non-GH interfaces, so I suppose that I should prefer to merge using the GH web UI in the future. For merging #239, I decided that I would rebase the relevant commits again myself. You are welcome to do this yourself or offer another course of action. Either way, I don't see what is so complicated – it is two |
Discussion in #223. Resolves #223.