On 9/15/2021 12:22 PM, Ken Brown wrote:
> One other small detail:
> 
> diff --git a/winsup/cygwin/select.cc b/winsup/cygwin/select.cc
> index 566cf66d6..cdb213a42 100644
> --- a/winsup/cygwin/select.cc
> +++ b/winsup/cygwin/select.cc
> @@ -645,7 +645,7 @@ pipe_data_available (int fd, fhandler_base *fh, HANDLE h, 
> bool writing)
>              return 1; /* We cannot know actual write pipe space. */
>            DWORD nbytes_in_pipe;
>            if (!PeekNamedPipe (query_hdl, NULL, 0, NULL, &nbytes_in_pipe, NULL))
> -           return 1;
> +           return -1;
>            fpli.WriteQuotaAvailable = fpli.InboundQuota - nbytes_in_pipe;
>          }
>         if (fpli.WriteQuotaAvailable > 0)
> 
> Having noticed that, I checked the callers of pipe_data_available with writing 
> == true, and they don't check for a return of -1.  So I think I need to fix that 
> too.

Could one of you (Corinna or Takashi) please check that I haven't done anything 
stupid in the attached patches?  If they look OK, I'll push them, and then maybe 
we're ready for a test release.

Ken