indygreg added a comment.
After coding up a handful of patches to switch to the new interface, I concede that I like the use of methods on the existing API. However, it is important for protocol version 2 that the peer interface switch to futures. I'm very reluctant to change the return value of peer APIs because that would likely be a difficult API break. I would be OK with exposing attributes on the command executor instance that magically route to commands. That's how the existing `iterbatch()` interface works. While the code is a bit wonky, it would work. REPOSITORY rHG Mercurial REVISION DETAIL https://phab.mercurial-scm.org/D3267 To: indygreg, #hg-reviewers Cc: mercurial-devel _______________________________________________ Mercurial-devel mailing list Mercurial-devel@mercurial-scm.org https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel
