9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
From: Ralph Corderoy <ralph@inputplus.co.uk>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] refresh problems with drawterm on x (freebsd)?
Date: Fri,  7 Mar 2003 16:19:35 +0000	[thread overview]
Message-ID: <454b.3e68c3c9.e9a0a@blake.inputplus.co.uk> (raw)
In-Reply-To: <18aefafa4a55e60cd64469860ad0c6a2@plan9.bell-labs.com>

Hi Russ,

> it seems pretty clear that painting the background is completing after
> painting the foreground characters, but the question is why.  we have
> the same problem with the accelerated hardware drivers -- drawing the
> background is an accelerated op so you have to issue it and then wait
> for it to complete before trying to draw the text.  i did not realize
> that sequential X calls can overlap like that, but empirically, they
> must.

I'm coming into the conversation part-way through here, but if you're
saying you've an X client doing XDrawText() and XFillRectangle() Xlib
calls, or you're generating the equivalent protocol, there's no way that
the operations should be done out of order.

> yesterday i tried inserting calls to XSync() after issuing the
> XFillRectangle calls, but that didn't help.  i still need to try
> XFlush.  i really wish the documentation were better.

X protocol is asynchronous, but the order is maintained.  You can watch
the protocol by putting a `snooper' in the middle, but if you find a
filled rectangle drawn before the text is covering it then it suggests
an X server bug, probably specific to the graphic card's driver.  A
mixture of accelerated and non-accelerated ops are the driver's problem,
not the X client's or the device independent parts of the X server.

Does this only occur with a specific X server?  If multiple people are
seeing it with different graphic cards/X servers, then it suggests a
client bug, possibly a dodgy Graphics Context results in the
text-drawing not changing any pixels.

Cheers,

-- 
Ralph Corderoy.      http://inputplus.co.uk/ralph/     http://troff.org/


  reply	other threads:[~2003-03-07 16:19 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-03-05 17:59 Russ Cox
2003-03-05 12:36 ` Philippe Anel
2003-03-05 18:46 ` andrey mirtchovski
2003-03-05 18:50   ` Dan Cross
2003-03-07  1:55     ` Kenji Arisawa
2003-03-07  6:14       ` andrey mirtchovski
2003-03-07 23:12       ` Russ Cox
2003-03-08  0:06         ` Geoff Collyer
2003-03-08  6:45       ` Kenji Arisawa
2003-03-13  9:55         ` Jeff Sickel
2003-03-13 12:11           ` Kenji Arisawa
2003-03-13 13:58             ` [9fans] Drawterm on MacOS X; weird colormap? Kenji Arisawa
2003-03-05 21:36   ` [9fans] refresh problems with drawterm on x (freebsd)? anyrhine
2003-03-05 20:18     ` Russ Cox
2003-03-05 21:55       ` anyrhine
2003-03-06 15:38         ` Philippe Anel
2003-03-06 14:08           ` Russ Cox
2003-03-06 14:16             ` Russ Cox
2003-03-07 16:19               ` Ralph Corderoy [this message]
2003-03-07 14:15                 ` Russ Cox
2003-03-10 10:18                   ` Ralph Corderoy

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=454b.3e68c3c9.e9a0a@blake.inputplus.co.uk \
    --to=ralph@inputplus.co.uk \
    --cc=9fans@cse.psu.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).