From mboxrd@z Thu Jan  1 00:00:00 1970
To: 9fans@cse.psu.edu
Subject: Re: [9fans] page and plumbing
From: "Russ Cox" <rsc@swtch.com>
Date: Sun, 30 Sep 2007 14:46:00 -0400
In-Reply-To: <20070926080547.GA4978@shodan.homeunix.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Message-Id: <20070930184603.9AFA81E8C22@holo.morphisms.net>
Topicbox-Message-UUID: c8561316-ead2-11e9-9d60-3106f5b1d025

> After a pull two days ago I noticed that plumbing images to page does not
> work any more. Page still starts but exits immediately before displaying
> anything. Starting page manually works as expected. As postscript and pdf
> files as well as other plumbing targets are unaffected led me to to suspect
> the new image handling code somehow interfers with plumbing.

True enough.  Here is a better solution:

cpu% diff -c /n/sources/plan9//sys/src/cmd/page/cache.c /sys/src/cmd/page/cache.c
/n/sources/plan9//sys/src/cmd/page/cache.c:151,156 - /sys/src/cmd/page/cache.c:151,159
  	Image *im;
  	int ra;

+ 	if(doc->npage < 1)
+ 		return display->white;
+
  	im = _cachedpage(doc, angle, page, "");
  	if(im == nil)
  		return nil;
cpu%

Sorry about that -- I'd forgotten about -i,
which has always been a hack.

I've submitted this as a patch.

It is unfortunately not related to
Greg Pavelcak's problems, which
I have not been able to reproduce.
The new page works fine for me in
drawterm.  That sounds like drawterm
is running out of image memory
somehow, but I don't see how, and
even if it did it should be behaving
more gracefully.

Russ