Good catch! I've fixed that (by commenting it out) in the GitHub asyncio repo.
I'm not going to bother right now with the CPython repo -- it's too much of a pain to merge three branches (and it's not strictly needed there), but this will happen the next time something gets pushed from the asyncio repo to CPython. On Fri, Nov 20, 2015 at 6:08 AM, Andrew Svetlov <[email protected]> wrote: > The patch breaks Python 3.3 test by using not existing yet > TestCase.subTest() call. > > On Thursday, November 19, 2015 at 11:40:55 PM UTC+2, Guido van Rossum wrote: >> >> Thanks everyone! This patch would have been a disaster without the >> recent feedback. Now I'm fairly confident that it works as intended, >> and I've committed it (with some doc changes too). >> >> On Thu, Nov 19, 2015 at 12:12 PM, Andrew Svetlov >> <[email protected]> wrote: >> > All tests passed, thanks! >> > >> > On Thu, Nov 19, 2015 at 10:06 PM, Guido van Rossum <[email protected]> >> > wrote: >> >> Thanks!! Try the new patch. >> >> >> >> I'm not keen on doing the warning any more given how simple it is to >> >> just make both idioms work. >> >> >> >> On Thu, Nov 19, 2015 at 11:28 AM, Andrew Svetlov >> >> <[email protected]> wrote: >> >>> The latest patch doesn't work also: it hangs on selector.select() call >> >>> if no network activity but loop._stopped == True >> >>> I suggest using zero timeout if loop._stopped is set. >> >>> >> >>> I agree that using loop.call_soon(loop.stop) is better but we already >> >>> have loop.stop() usage. >> >>> Removing the later sounds good but requires deprecation period IMHO >> >>> (with raising a warning as GvR suggested). >> >> >> >> >> >> >> >> -- >> >> --Guido van Rossum (python.org/~guido) >> > >> > >> > >> > -- >> > Thanks, >> > Andrew Svetlov >> >> >> >> -- >> --Guido van Rossum (python.org/~guido) -- --Guido van Rossum (python.org/~guido)
