sane-project-website/old-archive/1998-12/0067.html

<!-- received="Mon Dec  7 04:34:17 1998 PST" -->
<!-- sent="Mon, 07 Dec 1998 14:33:52 +0200" -->
<!-- name="Milon Firikis" -->
<!-- email="milonf@ariadne-t.gr" -->
<!-- subject="Re: SANE standard," -->
<!-- id="" -->
<!-- inreplyto="SANE standard," -->
<title>sane-devel: Re: SANE standard,</title>
<h1>Re: SANE standard,</h1>
<b>Milon Firikis</b> (<a href="mailto:milonf@ariadne-t.gr"><i>milonf@ariadne-t.gr</i></a>)<br>
<i>Mon, 07 Dec 1998 14:33:52 +0200</i>
<p>
<ul>
<li> <b>Messages sorted by:</b> <a href="date.html#67">[ date ]</a><a href="index.html#67">[ thread ]</a><a href="subject.html#67">[ subject ]</a><a href="author.html#67">[ author ]</a>
<!-- next="start" -->
<li> <b>Next message:</b> <a href="0068.html">Milon Firikis: "Re: USB scanners?"</a>
<li> <b>Previous message:</b> <a href="0066.html">Kaare Rasmussen: "HP1100A ?"</a>
<li> <b>Maybe in reply to:</b> <a href="0042.html">Robert Schwebel: "SANE standard,"</a>
<!-- nextthread="start" -->
<li> <b>Next in thread:</b> <a href="0071.html">Robert Schwebel: "Re: SANE standard,"</a>
<!-- reply="end" -->
</ul>
<!-- body="start" -->
Robert Schwebel wrote:<br>
<i>&gt; </i><br>
<i>&gt; Hello!</i><br>
<p>
[snip]<br>
<p>
<i>&gt; </i><br>
<i>&gt; A few questions:</i><br>
<i>&gt; </i><br>
<i>&gt; - Is there more information about how to write a backend? I looked threw</i><br>
<i>&gt;   the SANE coding standard V1.0 and found it still difficult to write a</i><br>
<i>&gt;   backend myself with what I found there. I played around with the pnm</i><br>
<i>&gt;   driver which seems to be a sample implementation, but it is very</i><br>
<i>&gt; poorly</i><br>
<i>&gt;   documented and other drivers seem to follow completely different</i><br>
<i>&gt;   strategies and style guides. Something like a "How to start writing a</i><br>
<i>&gt;   SANE backend" would be nice.</i><br>
<p>
That's the only thing AFAIK. However you did the right moves. PNM<br>
backend is difficult to understand and the scanners backend are way too<br>
different from your needs but still is an invaluable "source" of<br>
knowledge. I recommend mustek.c I started with that. You can also try<br>
apple.c. apple.c has a define<br>
#if 0<br>
#define NEUTRALIZE_BACKEND<br>
#define APPLE_MODEL_SELECT APPLESCANNER<br>
#endif<br>
<p>
in the apple.h. If you turn the zero to one then you will be able to use<br>
the backend even without scanner. You can't do anything useful with it<br>
but it will let you see the options handling and all these...<br>
<p>
<i>&gt; </i><br>
<i>&gt; - How do I compile my own backend? At the moment I copied my source into</i><br>
<i>&gt;   the sane-1.0/backends directory as root, compiled it there and</i><br>
<i>&gt; installed</i><br>
<i>&gt;   the libraries by hand in /usr/local/lib/sane/ (because make install</i><br>
<i>&gt;   installs always all libraries found, no matter if they already were</i><br>
<i>&gt;   installed). Is this really the right way? Any possibility of making</i><br>
<i>&gt;   the source with user permissions and only installing it as root?</i><br>
<p>
untar the SANE source<br>
chown -R user.group sane-XXX<br>
always edit source code as user (not root)<br>
always do make as user<br>
If you are a developer you should try ./configure with --static-libs or<br>
something, otherwise you won't be able to debug your backend.<br>
<p>
When you want to install, you install as root.<br>
<p>
In order to make your backend part of the building procedure you should<br>
edit Makefile.in and add your backend to the proper places.<br>
<p>
<i>&gt; - A few samples of how to deal with all these options defined in the</i><br>
<i>&gt; option</i><br>
<i>&gt;   descriptor would be nice as well. I tried to setup a string_list, but</i><br>
<i>&gt;   if I run xscanimage and select one of the list entries it stops with</i><br>
<i>&gt;   xscanimage: can't resolve symbol 'sanei_constrain_value'</i><br>
<p>
Give more info... <br>
<p>
<i>&gt; - Is there a separation between the API part of the SANE backend and the</i><br>
<i>&gt;   hardware dependend routines?</i><br>
<p>
SANE hides out scsi interface. You should use sanei_scsi routines and<br>
not direct scsi calls. Unfortunately this isn't true for other kind of<br>
interfaces. Especially for the scanners that have a proprietary ISA card<br>
one should use direct hardware port access giving up both compatibility<br>
(PC only) and security (SANE should be setuid root). For the parallel<br>
ports and serial ports I don't have a clear idea what is the status qwo<br>
right now. Maybe somebody else could clarify further?<br>
<p>
<i>&gt;Am I simply allowed to access hardware</i><br>
<i>&gt;   ports from my backend and are there any guidelines dealing with</i><br>
<i>&gt;   compatiblity (what system libraries am I allowed to use) and</i><br>
<i>&gt; portability?</i><br>
<p>
Give some more info (Hardware, interface, libraries) for a more proper<br>
advise and not a general conversation...<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="0068.html">Milon Firikis: "Re: USB scanners?"</a>
<li> <b>Previous message:</b> <a href="0066.html">Kaare Rasmussen: "HP1100A ?"</a>
<li> <b>Maybe in reply to:</b> <a href="0042.html">Robert Schwebel: "SANE standard,"</a>
<!-- nextthread="start" -->
<li> <b>Next in thread:</b> <a href="0071.html">Robert Schwebel: "Re: SANE standard,"</a>
<!-- reply="end" -->
</ul>