[sane-devel] Backend calibration data storage...
On Tue, Aug 19, 2003 at 02:39:37PM +0200, Gerhard Jaeger wrote: Hi there, I remember, that there once was a discussion 'bout saving backend specific data somewhere, but I could not find this discussion anymore. ... this will be my fall-back solution, but the better one will be to save the calibration data on on the disc... ... Hi, For the canon_pp backend we received two types of calibration information: a gamma curve, the format of which we didn't quite understand (but it works good if you just feed it back to the scanner later), and light dark current readings for the LIDE LED/sensor combo. Without the light/dark current value, the scanners produce nasty streaky images. Without the gamma curve the scanners produce blue ugly or yellow images, depending on the model. The backend saves both pieces of calibration information in a configurable location which defaults to ~/.sane/canon_pp-calibration-portname if no setting is given. The setting can be changed on a per-port basis in canon_pp.conf. I don't mind changing that if need be, but I think we should try and be consistant with this. I decided that making a configurable calibration file was most useful because it allowed a multi-user system to take advantage of only having to calibrate once. The driver should be robust in loading this file and fall-back to calibrating in memory, although it's a rather slow process so users will want to avoid it if possible. Cheers, - Matthew
iscan compile problems on Mandrake 9.1 (was Re: [sane-devel] (no subject))
On Tue, Aug 19, 2003 at 02:23:10PM +0200, flori...@gmx.net wrote: Hello! I have problems compiling iscan 1.5.2! What sources did you download? What am I missing? My system is mandrake 9.1 sane is compiled (Version 1.0.12). You'll need the gcc-3.x experimental sources. Either the tarball or the SRPM. Thanks Florian! make[1]: Entering directory `/root/iscan-1.5.2/frontend' Gack!! Am I right in thinking that you build your sources are root in root's home directory? You're better of building as a normal user and only switch to root for the installation. The rest of the log shows messages that are indicative of trying to build iscan from the gcc-2.9x sources on a system that uses gcc-3.x for its default compiler. HTH, -- Olaf MeeuwissenEPSON KOWA Corporation, ECS GnuPG key: 6BE37D90/AB6B 0D1F 99E7 1BF5 EB97 976A 16C7 F27D 6BE3 7D90 Penguin's lib! -- I hack, therefore I am -- LPIC-2
[sane-devel] Re: Re: My details
This is a multipart message in MIME format --_NextPart_000_38155476 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 7bit See the attached file for details --_NextPart_000_38155476--
[sane-devel] question about Option_Value union
Hello, I noticed that each backend defines an union like that: typedef union { SANE_Word w; SANE_Word *wa; SANE_String s; } Option_Value; Sometimes with fewer fields, sometime with more. Would it be a good idea to have it declared once for all backends in only one include file ? Regards, Stef
[sane-devel] Backend calibration data storage...
Hi, On Tue, Aug 19, 2003 at 02:39:37PM +0200, Gerhard Jaeger wrote: I remember, that there once was a discussion 'bout saving backend specific data somewhere, but I could not find this discussion anymore. My problem: I'd like to save calibration data from the devices locally, so that it is no longer necessary to redo calibration over and over again (at least for raw calibration this is always a time-intensive procedure). One solution is to keep this data only until the backend has been closed, this will be my fall-back solution, but the better one will be to save the calibration data on on the disc... Any comments? I'd use ~/.sane/some_unique_name, or, if you need more than one file, ~/.sane/backend_name/some_name. Better don't use /tmp, at least if you can't make sure that someone else can do nasty things with your files (e.g. symlink attack). Further more in /tmp, the file might be deleted on the next boot. Using ~/.sane means that the calibration is user-specific but I think that's better than causing security nightmares in /tmp. Bye, Henning
[sane-devel] Backend calibration data storage...
On Wed, Aug 20, 2003 at 09:58:51AM +0200, Henning Meier-Geinitz wrote: I'd use ~/.sane/some_unique_name, or, if you need more than one file, ~/.sane/backend_name/some_name. Better don't use /tmp, at least if you can't make sure that someone else can do nasty things with your files (e.g. symlink attack). Further more in /tmp, the file might be deleted on the next boot. I would hope that you're calibrating more often than every boot. My machines only boot a few times a year... Monty
[sane-devel] Backend calibration data storage...
On Mittwoch, 20. August 2003 00:33, Matthew Duggan wrote: [SNIPSNAP] Hi, For the canon_pp backend we received two types of calibration information: a gamma curve, the format of which we didn't quite understand (but it works good if you just feed it back to the scanner later), and light dark current readings for the LIDE LED/sensor combo. Good hint, 'll check the code Without the light/dark current value, the scanners produce nasty streak= y images. Without the gamma curve the scanners produce blue ugly or yellow images, depending on the model. The backend saves both pieces of calibration information in a configura= ble location which defaults to ~/.sane/canon_pp-calibration-portname if no setting is given. The setting can be changed on a per-port basis in canon_pp.conf. I don't mind changing that if need be, but I think we should try and be consistant with this. I decided that making a configurable calibration file was most useful because it allowed a multi-user system to take advantage of only having to calibrate once. Good idea... The driver should be robust in loading this file and fall-back to calibrating in memory, although it's a rather slow process so users will want to avoid it if possible. Well, the fall-back will be the current default behaviour: recalibration = with each scan ;-) Thanks for all the replies... Gerhard
[Fwd: Re: [sane-devel] ICC Profiler Comparison]
On Wed, 2003-08-06 at 00:38, Karl Heinz Kremer wrote: I had to change the -lqt linker parameter to -lqt-mt in order to get working profilers. The Makefiles do use the environment variable QTDIR, so as long as this variable is set correctly, everything should compile without any problems. What error messages are you getting? Karl Heinz On Tuesday, August 5, 2003, at 02:15 PM, gerard klaver wrote: From: gerard klaver ger...@gkall.hobby.nl Date: Tue Aug 5, 2003 2:10:35 PM America/New_York To: k...@khk.net Subject: Re: [sane-devel] ICC Profiler Comparison Reply-To: ger...@gkall.hobby.nl On Tue, 2003-08-05 at 15:02, k...@khk.net wrote: FYI: This link was just posted to the LCMS mailing list: http://www.tkupfer.de/imaging/Scan_Profiling.html It compares different profiler that can create ICC profiles for scanners. LCMS looks pretty good. ___ Sane-devel mailing list sane-de...@www.mostang.com http://www.mostang.com/mailman/listinfo/sane-devel Nice report, the LCMS windows version also works with Wine on linux. The linux version not yet used/compiled, some qt dependency/path problems. If somebody has it working i am interesting when modifications are made in the source especially when it are modifications for Debian (i use testing/unstable) -- -- m.vr.gr. Gerard Klaver Thanks, Problems solved with the following: lqt added -mt /usr path added where needed (QTDIR) solved dependency problems: mix up from qt3, qt2 files qt changed to qt3 in path names For whom are interested on http://www.bearteam.org/debian/unstable/ there are some .deb files from lprof-1.08 package difference with lprof-1.09? -- -- m.vr.gr. Gerard Klaver