Re: cygwin 3.3.x: another problem that may be related to pipes

[Takashi Yano <takashi.yano@nifty.ne.jp>]

On Mon, 15 Nov 2021 14:01:12 +0100
Corinna Vinschen wrote:
> [Redirected to cygwin-developers]
> 
> On Nov 15 17:18, Takashi Yano via Cygwin wrote:
> > On Sun, 14 Nov 2021 20:50:59 +0000 (UTC)
> > bf wrote:
> > > I've a shell script that processes several hundred xz-compressed text files with a pipeline of the form:
> > > 
> > > find . -depth -type f  ... -print0 | \
> > > xargs -0tI @ xzcat -- @ | \
> > > iconv --byte-subst='�' --unicode-subst='�' --widechar-subst='�' -f MS-ANSI -t UTF-8 -- | \
> > > grep ... \
> > > sed ... | \
> > > sort ... | \
> > > sort ...
> > > 
> > > . Since updating to cygwin 3.3.x (but not on cygwin 3.2.x) the script consistently fails when decompressing a portion of the files, resulting in lost data, and error messages of the form:
> > > 
> > > xzcat: (stdout): Write error: Bad address
> > 
> > Thanks for the report.
> > I could reproduce your problem and found the cause.
> > 
> > raw_read()/raw_write() in fhandler_pipe.cc seems to need handling
> > of STATUS_PENDING in nonblocking mode.
> > 
> > I also confirmed the following patch fixes the issue. However, I
> > am not very sure that this is the right thing.
> > 
> > Corinna, Ken, what do you think?
> 
> I'm puzzled.  non-blocking pipes return STATUS_PENDING?  What on earth
> is that supposed to mean on non-blocking (as opposed to asynchronous)
> pipes?  The problem is that STATUS_PENDING theoretically requires
> to wait for... something, the pipe handle, perhaps, and then to check
> io.Status.
> 
> Is it possible that nonblocking pipes need to be synchronous, i. e.,
> setting the FILE_SYNCHRONOUS_IO_NONALERT would fix that?

Yes. Setting FILE_SYNCHRONOUS_IO_NONALERT for the write pipe seems
to fix the issue. Proberbly, this needs the modification for
raw_write() like the modification I have made for raw_read() in
topic/pipe branch.

-- 
Takashi Yano <takashi.yano@nifty.ne.jp>