kopia lustrzana https://gitlab.com/sane-project/website
71 wiersze
3.3 KiB
HTML
71 wiersze
3.3 KiB
HTML
<!-- received="Mon Mar 1 03:16:02 1999 PST" -->
|
|
<!-- sent="Mon, 01 Mar 1999 13:14:38 +0200" -->
|
|
<!-- name="Milon Firikis" -->
|
|
<!-- email="milonf@ariadne-t.gr" -->
|
|
<!-- subject="Re: Porting SANE backends" -->
|
|
<!-- id="" -->
|
|
<!-- inreplyto="Porting SANE backends" -->
|
|
<title>sane-devel: Re: Porting SANE backends</title>
|
|
<h1>Re: Porting SANE backends</h1>
|
|
<b>Milon Firikis</b> (<a href="mailto:milonf@ariadne-t.gr"><i>milonf@ariadne-t.gr</i></a>)<br>
|
|
<i>Mon, 01 Mar 1999 13:14:38 +0200</i>
|
|
<p>
|
|
<ul>
|
|
<li> <b>Messages sorted by:</b> <a href="date.html#1">[ date ]</a><a href="index.html#1">[ thread ]</a><a href="subject.html#1">[ subject ]</a><a href="author.html#1">[ author ]</a>
|
|
<!-- next="start" -->
|
|
<li> <b>Next message:</b> <a href="0002.html">Milon Firikis: "Re: SCSI not detected"</a>
|
|
<li> <b>Previous message:</b> <a href="0000.html">Dr. Andreas Bolsch: "Re: Mustek ScanExpress 12000SP"</a>
|
|
<!-- nextthread="start" -->
|
|
<!-- reply="end" -->
|
|
</ul>
|
|
<!-- body="start" -->
|
|
David Mosberger-Tang wrote:<br>
|
|
<i>> </i><br>
|
|
<i>> >>>>> On Sun, 28 Feb 1999 18:21:32 +0200, Milon Firikis <<a href="mailto:milonf@ariadne-t.gr">milonf@ariadne-t.gr</a>> said:</i><br>
|
|
<i>> </i><br>
|
|
<i>> Milon> Note that the objective is to hide all Unix specific</i><br>
|
|
<i>> Milon> functions from the backend much like the sanei_scsi_ routines</i><br>
|
|
<i>> Milon> do for scsi operations.</i><br>
|
|
<i>> </i><br>
|
|
<i>> You're wrong. Yes, the _API_ must be OS independent but the sample</i><br>
|
|
<i>> implementation that we distribute is developed and intended for</i><br>
|
|
<i>> POSIX-compatible systems. If we can port that source to other</i><br>
|
|
<i>> platforms easily and cleanly, great, but to say that we want to hide</i><br>
|
|
<i>> all UNIX/POSIX specific functions is silly.</i><br>
|
|
<p>
|
|
Since the backends are mostly mid level drivers they should contain all<br>
|
|
the scanner protocol information just as they do now. I just thought<br>
|
|
that it would be nicer if the InterPorcess communication was somehow<br>
|
|
(Thread | Process) handled by the SANE API and not by the backend<br>
|
|
itself.<br>
|
|
<p>
|
|
<i>> The issue with fork() is that it's hard (inefficient) to implement on</i><br>
|
|
<i>> any system that doesn't have virtual memory. However, threads do not</i><br>
|
|
<i>> necessarily solve the problem either. E.g., on systems supporting</i><br>
|
|
<i>> user-level threads only (e.g., SunOS, HP-UX up to 10.20, etc.), a</i><br>
|
|
<i>> thread may not even help for SANE backends.</i><br>
|
|
<i>> </i><br>
|
|
<p>
|
|
I thought OS/2 had virtual memory management. I am not sure though.<br>
|
|
<p>
|
|
User level threads are out of the question if I understood you<br>
|
|
correctly. You need kernel level threads in order to allow non blocking<br>
|
|
access to frontend.<br>
|
|
<p>
|
|
MF<br>
|
|
<p>
|
|
<pre>
|
|
--
|
|
Source code, list archive, and docs: <a href="http://www.mostang.com/sane/">http://www.mostang.com/sane/</a>
|
|
To unsubscribe: echo unsubscribe sane-devel | mail <a href="mailto:majordomo@mostang.com">majordomo@mostang.com</a>
|
|
</pre>
|
|
<!-- body="end" -->
|
|
<p>
|
|
<ul>
|
|
<!-- next="start" -->
|
|
<li> <b>Next message:</b> <a href="0002.html">Milon Firikis: "Re: SCSI not detected"</a>
|
|
<li> <b>Previous message:</b> <a href="0000.html">Dr. Andreas Bolsch: "Re: Mustek ScanExpress 12000SP"</a>
|
|
<!-- nextthread="start" -->
|
|
<!-- reply="end" -->
|
|
</ul>
|