[sane-devel] SANE2 plans: RFC
Le Samedi 9 Ao=FBt 2003 12:34, Major A a =E9crit : Hi, I think that we should get the SANE2 implementation rolling, even if the standard isn't completely finalized. A standard is never final without even a single frontend and backend to use it. As Henning has said, adapting the test backend and scanimage should be first. But how shall we do that in CVS? The frontend should be easy enough because it can check for the backend version to see whether SANE1 or SANE2 is used, making it completely transparent to the user. But there is no version-check facility for frontends, so backends will be a bit more difficult.=20 Just a thought. Could SANE2 introduce a handcheck protocol to determine whi= ch=20 API is used : frontend_knows_versions( list_of_versions ) backend_agrees_to_use( version ) frontend_confirms_use_version( version ) If those functions are not found in backend - SANE1 If those functions are not called by frontend - SANE1 That would avoid a CVS mess up. As I did not follow the whole thread about that, I may be completely wrong,= =20 tell me if this is the case. Have a nice day, Thomas.
backend status: (was Re:[sane-devel] Current TODO list)
Thank you Henning. The hp5400.desc should be compliant now. Have a nice day. Le Lundi 28 Juillet 2003 20:33, Henning Meier-Geinitz a =E9crit : Hi, On Mon, Jul 28, 2003 at 07:16:14PM +0200, Thomas Soumarmon wrote: in the hp5400.desc file, when I set the backend status to beta, the make gives me the following warning : [sane-desc] Warning: DEPRECATED backend status `beta': setting status of backend `hp5400' to `good' So i set it to good and then it says to me : [sane-desc] ERROR: unknown status of backend `hp5400': `:good' That's simple: there is no backend status anymore, only per-scanner statuses (stati? statii? whatever). So just remove that one. If you really want to mention that your backend is beta, you can use tha backend's comment header. Bye, Henning ___ Sane-devel mailing list sane-de...@www.mostang.com http://www.mostang.com/mailman/listinfo/sane-devel
[sane-devel] Current TODO list
Hi, TODO (2003-07-25) todo [snip] backends * All backends: Check that global variables are initialized propperly. sane_init can be called more than once (if sane_exit was called before)! Global variables must be either initialized in sane_init or reset in sane_exit. At least these backends segfault: - microtek - hp5400 Some work has been done for hp5400. Just a few questions : * may the backend be called after a sane_exit() ? * is this needed for const global variables ? * could you explain me what the problem is with global initialization ? doc --- * desc files: - Change all description files to use the new status keyword: avision, canon, canon630u, coolscan2, fujitsu, hp, hp5400, hpsj5s, microtek, microtek2, plustek, sharp, st400, umax1220u. done for hp5400.
backend status: (was Re:[sane-devel] Current TODO list)
Hi, Thank you Henning for your explanation. I thought it could be due to a platform specific weird initialization. in the hp5400.desc file, when I set the backend status to beta, the make gives me the following warning : [sane-desc] Warning: DEPRECATED backend status `beta': setting status of backend `hp5400' to `good' So i set it to good and then it says to me : [sane-desc] ERROR: unknown status of backend `hp5400': `:good' Bye, Thomas.
[sane-devel] error in linux/videodev.h
Hi, snip Follows the errors : In file included from /usr/include/linux/videodev.h:14, from v4l.c:76: /usr/include/linux/videodev2.h:432: parse error before v4l2_std_id /usr/include/linux/videodev2.h:432: ISO C forbids data definition with no type or storage class Maybe that's the result over our pedantic warning settings. Try with --disable-warnings. Looks like the header file is not ISO C-conforming (for whatever version of ISO-C). I'll try that later and tell you what it changes. Where did you get videodev2.h from? Is it from your version of libc (which one)? Or is /usr/include/linux a link to /usr/src/linux/include/linux and you have installed a 2.5 kernel? Looking with the rpmdrake tool (Mdk 9.1 installed) it says that videodev2.h belongs to : glibc-devel-2.3.1-10mdk or kernel-source-2.4.21-0.13mdk I am wondering if this is exact or if it belongs to another package I would have tried on the machine (they have been numerous). We'll have to think about supporting v4l version 2 anyway. There is an old v4l2 backend, but nobody yet responded to the question if v4l2 is the same as the new version 2 kernel interface. Have a nice day, Thomas.
[sane-devel] Please test: changed handling of list of backends
Hi, I tested it with BACKENDS=hp5400 and it works great. It is nice to be able to compile only one backend without changing Makefile.in. I still have a problem compiling linux/videodev.h that prevents me from compiling v4l. I send another mail to talk about it. snip all
[sane-devel] error in linux/videodev.h
Hi, I have a problem compiling v4l because of the linux/videodev.h. Has anybody already had this problem with this file ? Follows the errors : In file included from /usr/include/linux/videodev.h:14, from v4l.c:76: /usr/include/linux/videodev2.h:432: parse error before v4l2_std_id /usr/include/linux/videodev2.h:432: ISO C forbids data definition with no type or storage class /usr/include/linux/videodev2.h:500: parse error before v4l2_std_id /usr/include/linux/videodev2.h:505: parse error before '}' token /usr/include/linux/videodev2.h:518: parse error before v4l2_std_id /usr/include/linux/videodev2.h:521: parse error before '}' token /usr/include/linux/videodev2.h:555: parse error before v4l2_std_id /usr/include/linux/videodev2.h:557: parse error before '}' token v4l.c: In function `sane_v4l_exit': Thank you for any help, Thomas.
[sane-devel] CVS freeze ?
Hi everybody, Sorry for the silly question but I have been a bit SANE-offline for a while, and I didn't find quickly the answer in the mail archive. I wanted to commit into CVS some changes about the hp5400 backend. Is there a CVS freeze at the moment or may I do it ? And before that, is there a compile farm or some kind of, to verify at least it compiles ok on various platforms before commiting it ? Thomas. PS : changes are not functional, they are only to make a synchronization between hp5400backend sourceforge CVS project and SANE CVS project. PS2: I have been messing around a lot before finding the DEBUG_NOT_STATIC trick to avoid the missing symbol sane_debug_hp5400_call !!!
[sane-devel] backends 1.0.12 compile-error
Maybe on a linux before version 2.4.19 the structures sockaddr_storage in_addr do not have ss_family member. A solution to that would be to migrate to a newer kernel or compile an older version of SANE ? snip
[sane-devel] Last call for new backends
Hi Henning, We had about 6 beta testers report. Some of them had problems for=20 installation, but once installed, none reported big bugs. All of them installed on Linux. Another point is that Martijn will be out for 10 days, and I won't be abl= e to=20 help for an undetermined amount of time. I would be in favour to add the backend to the release as it looks to wor= k OK=20 on linux and therefore be useful to some people. And maybe, adding it to = the=20 release would make more developpers help us in our effort on the backend. A good thing would be to have it compiled only on Linux or explicitly at= the=20 risk of the user. Is there an option like --enable-hp5400 to the configu= re=20 script ? I think you have all the points to decide whether to add it or not. Maybe Martijn has more things to say ? Thank you for your great work, Thomas. Le Mercredi 16 Avril 2003 17:30, Henning Meier-Geinitz a =E9crit : Hi, On Thu, Apr 17, 2003 at 12:10:35AM +1000, Martijn van Oosterhout wrote: Well, the code compiles and seems to run. Unfortunatly I'm on holiday= for the next 10 days so you won't hear from me for a while. Ok. So should I add the backend in its current state? There is still enough time to fix bugs until release. Bye, Henning
[sane-devel] Last call for new backends
The hp5400 backend works ok with sane 1.0.11. The main functionality it=20 misses, is the 2400 dpi scan. We have not been working on it for one mont= h=20 due to lack a of spare time. SANE API I did not follow the changes that were made to the SANE API, if they exis= t, so=20 I don't know what the behaviour of the backend could be with 1.0.12 Transport protocol The last point I think of is that the communication with the scanner is m= ade=20 only through ioctl() function. To be SANE compliant, I suppose it would n= eed=20 to have more options than this one. I suppose this should not be a lot of= =20 work, but it is still too much for my own agenda. If someone would have time to integrate it into sane, it would be nice.=20 Integration would consist in : * port hp5400 Makefile into SANE compilation structure * verify hp5400 API is correct for SANE * support of others transport protocol=20 The latest source code is accessible via cvs at sourceforge : cvs -d:pserver:anonym...@cvs.sourceforge.net:/cvsroot/hp5400backend login= =20 cvs -d:pserver:anonym...@cvs.sourceforge.net:/cvsroot/hp5400backend co =20 hp5400/hp5400 I would be very pleased if someone would help us for the integration of t= he=20 hp5400 backend. Thank you, Thomas. Le Lundi 14 Avril 2003 20:18, Henning Meier-Geinitz a =E9crit : Hi, On Mon, Apr 14, 2003 at 07:48:14PM +0200, Till Kamppeter wrote: What about also including these backends (they are listed in sane/sane-backends/doc/descriptions-external/ in the current CVS snapshot): Noone of the maintainers of these backends asked for inckusion as far as I know. I hesitate to add a backend without an active maintainer. hp5400 No idea. ibm Not much used I guess at least the feedback I got was pretty minimal. But maybe the author got more. I'll have a look if it can be included without too much changes. lhii The webpage is gone and I haven't heard about this for years. Anyone using handscanners anymore? niash No idea. onetouch8600 I'm getting a connection closed or document contains no data from the server currently. primax That one may need some work to integrate it into sane as it currently compiles stand-alone. scanwit Seems to think the device is detected without checking? v4l2 Is v4l2 used at all? viceo IIRC, can't be added because of license trouble. Bye, Henning ___ Sane-devel mailing list sane-de...@www.mostang.com http://www.mostang.com/mailman/listinfo/sane-devel
[sane-devel] Last call for new backends
snip What about libieee1284? I have seen some defines for that? Does the backend work with parport also? If yes, that'd need some work to only use libieee1284 if it's available. backend does not work with parport. I don't know if the communication pro= tocol=20 would fit anyway. The defines you saw are remains from the template backe= nd=20 used. Bye, Thomas.
[sane-devel] New release of the alpha version of hp5400 SANE backend
Hi everybody, Martijn have done more great work those days. It is still an alpha versio= n as=20 it is not full featured and there are still two major bugs :=20 * the image goes fuzzier as the resolution increases=20 * 1200 and 2400 dpi images are scrambled. Except of that, 75 and 150 dpi are working pretty well. The release is downloadable at http://sourceforge.net/projects/hp5400back= end/ The most significant modifications are : - improved calibration - implemented gamma correction - Add support for new SCANOFFSET command 0xE700. This means we can remove all the options relating to offsets in xsane itself. This might require adjustment depending on the model of scanner. - Add support for the two kinds of types of scan requests. This is what controls whether to use the scan offsets. - Add a ScanType parameter so we can distinguish between previews and nor= mal scans. In future we'll use this to control the scan speed. - Add code to protect the scanner against silly mistakes. It will restric= t your scan area to the glass and no more.=20 Have a nice day, Thomas.
[sane-devel] New release of the alpha version of hp5400 SANE backend
Martijn have done more great work those days. It is still an alpha ver= sion as it is not full featured and there are still two major bugs : * the image goes fuzzier as the resolution increases * 1200 and 2400 dpi images are scrambled. Except of that, 75 and 150 dpi are working pretty well. What do you mean by fuzzier? Is this the problem of vertical shift between separate color components (R G B)? Yes. That point is under control as we know how to correct it during th= e=20 image composition, reading colors on separate lines. Problems with scrambled images could be caused by an overflow of the internal data buffer (at least that's what I found with niash12/14/19 chips). In that case, you should be able to correlate this to backtrack= ing (quickly moving back and forth of the lamp). Do the scrambled images also occur if you scan only a narrow vertical strip? At first I thought you would be right but even with a narrow vertival str= ip,=20 most pixels are absent. I'll send you on a separate mail a tiny part of= the=20 vertical strip. It shows colored vertical lines (red and green). It might be related to the query sent to the scanner. We have to check if= all=20 the flags are correct. Thank you again for sharing your experience and for your help. Thomas.
[sane-devel] Problem installing new HP5400 backend
Hi Robe, Maybe you did not indicate in dll.conf that it should load hp5400 backend= ? edit /usr/local/etc/sane.d/dll.conf or /etc/sane.d/dll.conf=20 and on a new line add : hp5400=20 Note that direct load of backend (e.g. with -d hp5400) has not been teste= d=20 yet. Let us know if it works. Le Vendredi 28 F=E9vrier 2003 07:15, Rob Male a =E9crit : Hello, I have been trying to install the new HP5400 backend for my HP5470c on Redhat8.0. (but with no luck) I downloaded and installed sane-backends-1.0.11 and sane-find-scanner recognizes the scanner, however scanimage -L reports No scanners were identified I downloaded the HP5400 backend and modified the paths in the Makefile before doing a make install. I made no changes to the hp5400.conf file sane-find-scanner returns:- found USB scanner (vendor=3D0x03f0, product=3D0x1105) at /dev/usb/scan= ner0 # Your USB scanner was detected. It may or may not be supported by # SANE. Try scanimage -L and read the backend's manpage. cat on /proc/bus/usb/devices returns:- P: Vendor=3D03f0 ProdID=3D1105 Rev=3D 0.00 S: Manufacturer=3DHewlett-Packard S: Product=3DHP Scanjet 5400C Series I added the following line to /etc/rc.d/rc.local modprobe scanner vendor=3D0x03f0 product=3D0x1105 /usr/local/etc/sane.d/hp.conf has only 2 lines:- /dev/usb/scanner0 option connect-device Any suggestion on what to do now would be greatly appreciated as I have no idea where to start. Regards Rob Male ___ Sane-devel mailing list sane-de...@www.mostang.com http://www.mostang.com/mailman/listinfo/sane-devel
[sane-devel] Alpha version for hp5400 backend
Hi, Heh, the desc file was there, is just isn't in CVS. We have a 5400 and = a 5470c and the only difference we've seen so far is the version number returned. Maybe we need to check the HP site to see what the difference= s are. :) I think they are similar scanners just with more options : * 5470 =3D 5400 + XPA (for film rolls and diapositives) * 5490 =3D 5470 + ADF (auto feeder)
[sane-devel] Alpha version for hp5400 backend
I would like to add the .desc file to CVS (external backends) so your backend is found in the scanner search engine. I've edited your file a Oops, error. It should be in CVS. bit. If you write one entry for each model, it's found easier (nobody seaches for 54xx). Do you have a 5400 and 5470? As far as I know, the 5490 is similar, therefore I've added it as untested. Is this correct? Status is development under way with untested alpha release. I forgot to mention that it has only been compiled for the moment under l= inux=20 and can probably break on other OS.
[sane-devel] Linux 2.5.62
Couldn't it be the old problem of permissioning of /dev/usb/scanner1 ? Does the user has read and write access to the device file ? Le Mardi 25 F=E9vrier 2003 14:55, Stian Jordet a =E9crit : tir, 2003-02-25 kl. 14:26 skrev Henning Meier-Geinitz: Hi, Well, then I guess it's something else (not related to the kernel scanner driver). Please enable SANE_DEBUG_SANEI_USB=3D255 also. Ahh, this made more sense, I think (altough still no idea what the problem might be. Operasjonen er ikke tillatt is norwegian for something like Operation not permitted. [plustek] Checking for 0x04b8-0x011d [sanei_usb] sanei_usb_find_devices: vendor=3D0x04b8, product=3D0x011d, attach=3D0x405cd184 [plustek] Found device at libusb:002:003 [sanei_usb] sanei_usb_open: trying to open device `libusb:002:003' USB error: could not set config 1: Operasjonen er ikke tillatt [sanei_usb] sanei_usb_open: libusb complained: could not set config 1: Operasjonen er ikke tillatt [sanei_usb] Make sure you run as root or set appropriate permissions [plustek] open failed: -1 [plustek] sane_get_devices (0xbfffe848, 0) [plustek] sane_exit chevrolet:~# sane-find-scanner # No SCSI scanners found. If you expected something different, ma= ke sure that # you have loaded a SCSI driver for your SCSI adapter. found USB scanner (vendor=3D0x0731, product=3D0x0528) at libusb:001= :002 found USB scanner (vendor=3D0x04b8 [EPSON], product=3D0x011d [EPSON Scanner]) at libusb:002:003 What kind of device is the first one? The first device is a usb-cradle for my Sony Ericsson P800 phone... Thanks for your patience :) Best regards, Stian ___ Sane-devel mailing list sane-de...@www.mostang.com http://www.mostang.com/mailman/listinfo/sane-devel
[sane-devel] Alpha version for hp5400 backend
Hello everybody, Thanks to Martijn van Oosterhout, an alpha version of the SANE backend fo= r=20 HP5400 scanners is available for download at=20 http://sourceforge.net/projects/hp5400backend As it is alpha code, it is buggy and of interest only for development=20 purposes. Feel free to test it and submit bug fixes, improvements Have a nice day, Thomas.
[sane-devel] HP5470c scanner nearly working!
Le Jeudi 20 F=E9vrier 2003 13:30, Martijn van Oosterhout a =E9crit : Anyway, I've put a new version up on my site (http://svana.org/kleptog/hp5470/). Let me know if the images come out = OK. My xview is having problems showing 48bpp PPM images so you may have to convert them to PNG to make them viewable. Let me know how it goes. It compiled and ran perfectly. As in your output.png the scan area is a b= it=20 outside the glass and the colors are OK. I'll probably have time on next week to do some work on the scanner. Just= tell=20 me what you would like to do to help.=20 Great work.
[sane-devel] HP5470c scanner nearly working!
Hi Martijn, I began about one month ago to do the same work you've done. Unfortunatel= y=20 about 2 weeks after the beginning my spare time reduced a lot, so I must = say=20 that it slowed down a lot, reaching a 0 speed for about 1 month. There is now a Sourceforge project set up for hp5400 series SANE backend = :=20 http://sourceforge.net/projects/hp5400backend/ and=20 http://hp5400backend.sourceforge.net/ I hope you can get some information there. I think it would be worse you add your work on that project. As you might= have=20 more time than I, would you want to be added as an administrator on that=20 project ? Another thing. The test program you wrote compiled perfectly and appearen= tly=20 worked ok. I email you the log directly. Have a nice day. Le Mercredi 19 F=E9vrier 2003 23:52, Martijn van Oosterhout a =E9crit : [Please CC any replies so I can more easily find them through the rest = of my mail. Thanks.] I finally got totally sick of waiting for someone to make my scanner wo= rk, so I did it myself. The results are on my webpage: http://svana.org/kleptog/hp5470/. I'm doing it over USB. Situation is: 1. It finds the scanner and warms up the lamp 2. It performs two calibrations scans like in windows 3. Performs the preview scan and dumps it to disk This is only a test program, not a backend. Things still to do: 1. Actually use results of calibration scans. Anyone who has an idea ho= w scanners are calibrated, point me at some info, please. 2. Do gamma correction. Right now the gamma correction is linear. Again= , someone who knows the theory would be helpful here. 3. Decipher data format. Looks pretty straight forward. R, G, B seperat= e. 4. Work out the parameters for the different scan types (BW, 24-bit, e= tc). 5. Turn it into a SANE backend. Hopefully this will not be hard. 6. Make negative/slide addon work. Hopefully not far to go now. If anyone can test my code it would be muc= h appreciated. Note this version produces a *lot* of debug info. Please s= end it all for bug reports. Thanks. I'd like to thank Bertrik Sikken for doing some of the ground work. Have a nice day,
[sane-devel] sane-devel delivery slow?
Could the delay be due to the Slammer worm ? If the mailing server has a Microsoft SQL/Server it could be the case. I think of it because it began by the same date. Le Vendredi 31 Janvier 2003 09:21, Gerhard Jaeger a =E9crit : Hi, On Thursday 23 January 2003 11:22, Henning Meier-Geinitz wrote: Hi, Is everyone noticing a delay in sane-devel email arriving or is it just me? Mails send to the list arrive at mostang.com pretty quick (seconds to minutes) and show up in the archive. But they take up to 15 hours to reach me. E.g. I can see three answers to stupid question about SCSI in the archive but none has arrived here until now. For me, the delay occurs somewhere between mostang.com accepting the mails and sending them to the recipients. That could also be a problem with my provider but Abel D. told me about similar problems. Anyone else? in the meantime yes! Any explanations? Gerhard ___ Sane-devel mailing list sane-de...@www.mostang.com http://www.mostang.com/mailman/listinfo/sane-devel
[sane-devel] Generate C source code from a USBSnoopy log
Hi everybody, I have been playing around with my HP5470c scanner for some days without = any=20 good results. I tried to handwrite C code corresponding to USBSnoopy logs= I=20 got before as the logs are a bit long and tough, it never worked fully. S= o I=20 decided to replay the log with usb-robot. Logs did not fit with the expec= ted=20 syntax... annoying. So I decided to write an awk program that would write for me the C source= code=20 that replays the log, and it works. Now I'll need to add the code to write the image to a file. Could anybody= tell=20 me which format should I expect for the image (RGB pixels from top left ?= ).=20 Is there a library I should use for image saving and processing (ImageMag= ick=20 ?). Thank you for your help. I suppose that tha tool could be of interest for more people so it is=20 available for download at sourceforge : http://prdownloads.sourceforge.net/hp5400backend/USBSnoopy2CsourceCode.ta= r.gz?download Just a few notes : * This tool works with libusb. Feell free to port it to the kernel USB li= b if=20 you know how to do it. * you need awk and a C compiler to use it (if you don't have them, it is=20 probably of no use for you) Regs, Thomas.
[sane-devel] No devices available
The device you try to connect to should be accessible by the user who run= s=20 scanimage, xscanimage, xsane, etc... do a=20 ls -al /dev/scanner and ls -al /dev/sg1 to check that other users are allowed to read and write the device. If = not,=20 use chmod 766 /dev/sg1 chmod 766 /dev/scanner to grant read-write access to all users to those devices. =20 The other problem (connecting to display: :0.0) looks like a permissionin= g=20 issue under X.=20 You could change the DISPLAY environment variable before trying to run=20 xscanimage with root to set it to another display ? =20 Le Lundi 13 Janvier 2003 05:05, Barry Kirsten a =E9crit : Hi Henning, On Mon, 13 Jan 2003 08:15, Henning Meier-Geinitz wrote: Well, what does happen when you try to scan? Try scanimage image.pnm= =2E If that doesn't work, set the debug variable. Do the same for xsane. Do you use both xsane and scanimage as the same user? Otherwise there may be permission problems. I only one version of SANE installed? Otherwise xsane may use the other one. scanimage works, but only when run as root. When I try scanimage as a normal user I get the message: scanimage: no SANE devices found. This i= s similar to when I try to start xsane from KDE as normal user: xsane: no devices found. Note...sane and xsane are set up for anyone to execute, so not a permis= sion problem. There is certainly only one copy of xsane running. To be certain I trie= d removing it with apt-get remove twice, (and also for sane on its own) a= nd got nothing. I think this is part of the problem...when I try to start xsane as root= , I get: debian:/home/barry# xsane Xlib: connection to :0.0 refused by server Xlib: Client is not authorized to connect to Server Gtk-WARNING **: cannot open display: :0.0 I'm out of my depth here. Is this a problem with permissions/access somewhere in X? (Don't forget, I was root when starting xsane from terminal.) Also the version of xsane I'm using is 0.84-2 from the Debia= n stable CDs. Many thanks for your help, Barry ___ Sane-devel mailing list sane-de...@www.mostang.com http://www.mostang.com/mailman/listinfo/sane-devel
[sane-devel] backend for HP5470c (and probably HP5400 family)
Hello everybody, Owning an excellent HP scanjet 5470c scanner, I would like to make it wor= k=20 with my mandrake 9.0 linux box as I would like to get rid of Win2k. USB Snoopy helped me to get some logs of : * turn on/off the scanner * turn on/off the HP scanning software * preview scan in full color * scan whole area in Black White using these logs and the testtool upon libusb by Bertrik Sikken=20 (http://home.kabelfoon.nl/~bertrik/hp5400/index.html), I managed to make= the=20 scanner move but not to scan as some bulk_written data are missing in the= =20 logs and also because I have till now no experience with writing device=20 drivers. Finally, is there anyone out there who dedicates some time to write a bac= k-end for this family of scanner ? Could I help him/her in some way ? Else, is there a how-to about writing sane backend ? Happy new year to the sane family. Thomas