Small regression on Windows

Please type this, preferably outside of any repo to avoid that it is changed; note the command typo:

    git -c help.autocorrect=40 ceckout

Git waits for 4 seconds, but does not write anything until just before it exits. It bisects to

a9b8a09c3c30886c79133da9f48ef9f98c21c3b2 is the first bad commit
commit a9b8a09c3c30886c79133da9f48ef9f98c21c3b2
Author: Jeff Hostetler <jeffhost@microsoft.com>
Date:   Thu Dec 22 18:09:23 2016 +0100

     mingw: replace isatty() hack

     Git for Windows has carried a patch that depended on internals
     of MSVC runtime, but it does not work correctly with recent MSVC
     runtime. A replacement was written originally for compiling
     with VC++. The patch in this message is a backport of that
     replacement, and it also fixes the previous attempt to make
     isatty() tell that /dev/null is *not* an interactive terminal.

     Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
     Tested-by: Beat Bolli <dev+git@drbeat.li>
     Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
     Signed-off-by: Junio C Hamano <gitster@pobox.com>

:040000 040000 bc3c75bdfc766fe478e160a9452c31bfef50b505 f7183c3a0726fef7161d5f9ff8c997260025f1bb M compat

Any ideas? I haven't had time to dig any further.

-- Hannes

On 2/13/2017 1:00 PM, Johannes Sixt wrote:

I'm investigating now.

The warning text is being written to stderr and then main thread sleeps for the 4 seconds. However, stderr has been redirected to the pipe connected to the console_thread that handles the coloring/pager. It is in a blocking ReadFile on the pipe and is thus stuck until the main thread runs the corrected command and closes the pipe. It then sees the EOF and prints everything in the pipe buffer.

I'll look into fixing this now.

Jeff

Jeff Hostetler <git@jeffhostetler.com> writes:

IOW, somehow your stderr is not flushing automatically?

Thanks.