Yukihiro Nakadaira wrote:

> > > > > test_channel.vim fails with Win32 CUI Vim.
> > > > >
> > > > > What the following patch fix is
> > > > > 1. Vim sleep 5 msec even when timeout=0.
> > > > > 2. channel_handle_events() closes channel when there is no input.
> > > > > 3. ":sleep" command does not read channel (I am not sure where is the
> > > > right
> > > > >    place to call channel_handle_events()).
> > > >
> > > > Thanks!
> > > >
> > > > For Unix checking for events happens in mch_breakcheck().  And any
> > other
> > > > place where RealWaitForChar() is called.
> > > >
> > > > I hesitate to add channel_handle_events() in mch_breakcheck, it might
> > be
> > > > too slow.  Perhaps it's better to put it inside
> > parse_queued_messages().
> > > > That sort of makes sense.  And we can remove the other calls to
> > > > channel_handle_events() since parse_queued_messages() is called there
> > > > already.  They were actually in the wrong order.
> > > >
> > > > Let me make that change.  Please check if it works.
> > >
> > > It works fine.
> > >
> > >
> > > > Did you look at the problem that ch_read() doesn't timeout for Win32?
> > > > It's disabled in the test for now.
> > > >
> > >
> > > To implement timeout, we need to use select() instead of
> > WSAAsyncSelect().
> > > Patch is attached.  With this patch, Vim checks channel each 100
> > > milliseconds while waiting window message.
> >
> > Hmm, this completely removes the asynchronous handling of messages.
> > That is strange, I thought we needed that.
> >
> > Can't we have both asynchronous handling as soon as a message arrives,
> > and another way to wait for something to arrive in ch_read()?
> > I suppose that instead of using select() we would use pPeekMessage() and
> > loop with a short sleep in channel_wait(), like we do for pipes.
> >
> 
> I tried loop with PeekMessage() but I could not figure out how to
> implement it when there is multiple channel.
> 
> I also tried recv() with MSG_PEEK, but it did not work.
> 
> I think my patch is not so different because WM_NETBEANS is handled
> only when we call GetMessage().  Maybe we also need to call
> channel_handle_events() in gui_mch_update() to make it same with
> current behavior.
> 
> Perhaps we can use Overlapped I/O to wait both channel and message without
> delay.
> 
> Channel polling in the message loop is also required for pipe.  Because
> WSAAsyncSelect() can not handle pipe.

Thanks for the explanation.  Let me include this patch, at least it
makes one more test pass.

That 100 msec wait seems a bit long.  Perhaps we can make it short the
first time and longer when looping?

Let's see if there are more improvements we can make.

-- 
There can't be a crisis today, my schedule is already full.

 /// Bram Moolenaar -- [email protected] -- http://www.Moolenaar.net   \\\
///        sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org        ///
 \\\            help me help AIDS victims -- http://ICCF-Holland.org    ///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Raspunde prin e-mail lui