kopia lustrzana https://gitlab.com/sane-project/website
96 wiersze
4.4 KiB
HTML
96 wiersze
4.4 KiB
HTML
|
<!-- received="Sun Apr 5 14:00:17 1998 PDT" -->
|
|||
|
|
<!-- sent="Sun, 5 Apr 1998 14:00:16 -0700" -->
|
|||
|
|
<!-- name="David Mosberger-Tang" -->
|
|||
|
|
<!-- email="David.Mosberger@acm.org" -->
|
|||
|
|
<!-- subject="Re: sane 0.72pre meets FreeBSD" -->
|
|||
|
|
<!-- id="199804052100.OAA11604@panda.mostang.com" -->
|
|||
|
|
<!-- inreplyto="Pine.BSF.3.96.980405211747.306A-100000@atlas.home.dynas.se" -->
|
|||
|
|
<title>sane-devel: Re: sane 0.72pre meets FreeBSD</title>
|
|||
|
|
<h1>Re: sane 0.72pre meets FreeBSD</h1>
|
|||
|
|
<b>David Mosberger-Tang</b> (<a href="mailto:David.Mosberger@acm.org"><i>David.Mosberger@acm.org</i></a>)<br>
|
|||
|
|
<i>Sun, 5 Apr 1998 14:00:16 -0700</i>
|
|||
|
|
<p>
|
|||
|
|
<ul>
|
|||
|
|
<li> <b>Messages sorted by:</b> <a href="date.html#66">[ date ]</a><a href="index.html#66">[ thread ]</a><a href="subject.html#66">[ subject ]</a><a href="author.html#66">[ author ]</a>
|
|||
|
|
<!-- next="start" -->
|
|||
|
|
<li> <b>Next message:</b> <a href="0067.html">David Mosberger-Tang: "Re: I don<6F>t like that"</a>
|
|||
|
|
<li> <b>Previous message:</b> <a href="0065.html">John Brush: "Re: Q: support for digi cameras"</a>
|
|||
|
|
<!-- nextthread="start" -->
|
|||
|
|
<!-- reply="end" -->
|
|||
|
|
</ul>
|
|||
|
|
<!-- body="start" -->
|
|||
|
|
<i>>>>>> On Sun, 5 Apr 1998 21:21:08 +0200 (CEST), =?ISO-8859-1?Q?Mikko_Ty=F6l=E4j=E4rvi?= <<a href="mailto:mikko@securitydynamics.com">mikko@securitydynamics.com</a>> said:</i><br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> The patches below are all (?) that is required to compile and<br>
|
|||
|
|
Mikko> pass initial tests on FreeBSD 2.2.5 with gtk 0.99.9 and gimp<br>
|
|||
|
|
Mikko> 0.99.23.<br>
|
|||
|
|
<p>
|
|||
|
|
Thanks! I think I got all the fixes in place so hopefully 0.72 will<br>
|
|||
|
|
build out of the box.<br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> Question 0:<br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> Is the entire build supposed to pass without any compiler<br>
|
|||
|
|
Mikko> warnings?<br>
|
|||
|
|
<p>
|
|||
|
|
Yes and no. Yes if you have a modern enough libc that declares all<br>
|
|||
|
|
the necessary function, no in all other cases. "Modern enough" is<br>
|
|||
|
|
roughly equivalent to glibc 2.x.<br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> If so, there are a few things that could be improved:<br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> - getopt.c does not include headers to get declarations of<br>
|
|||
|
|
Mikko> getpid() and strdup(). Well, it sort of does, but there are<br>
|
|||
|
|
Mikko> enough #ifdefs to make sure it does not happen. But I guess<br>
|
|||
|
|
Mikko> this file is simply grabbed out of GNU libs, and should not<br>
|
|||
|
|
Mikko> be changed. No big deal.<br>
|
|||
|
|
<p>
|
|||
|
|
That's intentional. There are systems broken enough that _declare_<br>
|
|||
|
|
certain functions but do not _define_ them in any library (yes, this<br>
|
|||
|
|
is true). Because of this, it's safer not to declare replacement libc<br>
|
|||
|
|
functions at all. The warnings are annoying, but C's parameter<br>
|
|||
|
|
passing rules work in a way that makes this safer.<br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> - sane provides strndup() for systems that does not have it,<br>
|
|||
|
|
Mikko> but fails to declare it anywhere. It could be done in<br>
|
|||
|
|
Mikko> config.h, which would require to include <sys/types.h><br>
|
|||
|
|
Mikko> somewhere too (to get size_t).<br>
|
|||
|
|
<p>
|
|||
|
|
This, again is intentional for the above reason. glibc 2.x declares<br>
|
|||
|
|
and defines strndup() so there you won't see any warnings.<br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> Question 1:<br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> Is it necessary to keep all the loadable backends in the<br>
|
|||
|
|
Mikko> standard library path? I'm getting an uneasy feeling when<br>
|
|||
|
|
Mikko> sane singlehandedly doubles the number of entries in my<br>
|
|||
|
|
Mikko> /usr/local/lib, with something that could be considered as<br>
|
|||
|
|
Mikko> internal to sane.<br>
|
|||
|
|
<p>
|
|||
|
|
Mikko> Am I missing something, or could they be moved to (for<br>
|
|||
|
|
Mikko> example) /usr/local/lib/sane/? Aren't the backends just<br>
|
|||
|
|
Mikko> dlopen()ed?<br>
|
|||
|
|
<p>
|
|||
|
|
The probably should. There used to be few enough backends that this<br>
|
|||
|
|
didn't matter, but I think you're right that we're getting past this<br>
|
|||
|
|
point. I won't promise it for 0.72 but it's something we should<br>
|
|||
|
|
change (if someone were to send me tested patches, it would accelerate<br>
|
|||
|
|
things a lot, hint, hint... ;-)<br>
|
|||
|
|
<p>
|
|||
|
|
--david<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="0067.html">David Mosberger-Tang: "Re: I don<6F>t like that"</a>
|
|||
|
|
<li> <b>Previous message:</b> <a href="0065.html">John Brush: "Re: Q: support for digi cameras"</a>
|
|||
|
|
<!-- nextthread="start" -->
|
|||
|
|
<!-- reply="end" -->
|
|||
|
|
</ul>
|