[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Thu, 07 Dec 2006 21:34:15 +0100 (CET) Franz Bakan fba...@gmx.net wrote: On Thu, 7 Dec 2006 20:34:05 +0100, Alessandro Zummo wrote: I committed a new version that should be more portable. You may have to regenerate configure using autoconf. Hi Alessandro, now all compiles again here on OS/2 without errors. great! now tell me you have an epson scanner to test the driver with :) -- Best regards, Alessandro Zummo, Tower Technologies - Turin, Italy http://www.towertech.it
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Fri, 08 Dec 2006 17:37:31 +0100 (CET) Franz Bakan fba...@gmx.net wrote: On Fri, 8 Dec 2006 00:45:04 +0100, Alessandro Zummo wrote: now all compiles again here on OS/2 without errors. great! now tell me you have an epson scanner to test the driver with :) Well which epson scanners are supposed to work? there is no epson2.desc or epson2.man. I know, still have to write them. It is supposed to support the same scanners as epson backend. I have an epson GT-9500 I could try later this weekend. What's the benefit over epson backend? actually, only minor benefits. But I want to be sure I have not breaked compatibility with other scanners. In somewhat near future, epson2 will be able to scan using extended commands (they should be a little bit faster), talk to network capable scanners and to the FilmScan 200. I also would like to collect information on the scanners in order to maintain compatibility in the future. -- Best regards, Alessandro Zummo, Tower Technologies - Turin, Italy http://www.towertech.it
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Sat, 02 Dec 2006 21:26:25 +0100 (CET) Franz Bakan fba...@gmx.net wrote: On Sat, 2 Dec 2006 20:31:06 +0100, Alessandro Zummo wrote: well, you need to check where are those functions/types defined on os/2 :) Yes, I know and already checked, but unfortunattely there is not place where __le16_to_cpup(), __u16,, __le32_to_cpup(), __u32, __le32 or __cpu_to_le32() are defined on OS/2. What do you suggest? Hi Franz, I committed a new version that should be more portable. You may have to regenerate configure using autoconf. -- Best regards, Alessandro Zummo, Tower Technologies - Turin, Italy http://www.towertech.it
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Thu, 7 Dec 2006 20:34:05 +0100, Alessandro Zummo wrote: I committed a new version that should be more portable. You may have to regenerate configure using autoconf. Hi Alessandro, now all compiles again here on OS/2 without errors. Many thanks Franz
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Tue, 5 Dec 2006 09:12:43 + (GMT) Jon Chambers j...@jon.demon.co.uk wrote: Hi, On Tue, 5 Dec 2006, Gerhard Jaeger wrote: On Monday 04 December 2006 21:59, Alessandro Zummo wrote: BTW: In the Plustek backend, I use this macro #define _SWAP(x,y) { (x)^=(y); (x)^=((y)^=(x));} If they are suitable for you then htons, htonl, etc should be pretty ubiquitous across OS's. htons and htonl convert from host endian to big endian. I need little endian to host endian. the macro i used, which are common under linux, are evaluated a compile time . I will search for an appropriate macro in the next few days. -- Best regards, Alessandro Zummo, Tower Technologies - Turin, Italy http://www.towertech.it
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
Hi, On Tue, 5 Dec 2006, Gerhard Jaeger wrote: On Monday 04 December 2006 21:59, Alessandro Zummo wrote: BTW: In the Plustek backend, I use this macro #define _SWAP(x,y) { (x)^=(y); (x)^=((y)^=(x));} If they are suitable for you then htons, htonl, etc should be pretty ubiquitous across OS's. cheers, Jon == Jon Chambers = http://www.jon.demon.co.uk, 020 8575 7097, 07931 961669 =
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Saturday 02 December 2006 22:11, Alessandro Zummo wrote: On Sat, 02 Dec 2006 21:26:25 +0100 (CET) Franz Bakan fba...@gmx.net wrote: On Sat, 2 Dec 2006 20:31:06 +0100, Alessandro Zummo wrote: well, you need to check where are those functions/types defined on os/2 :) Yes, I know and already checked, but unfortunattely there is not place where __le16_to_cpup(), __u16,, __le32_to_cpup(), __u32, __le32 or __cpu_to_le32() are defined on OS/2. What do you suggest? mmm.. you should try to find similar macros we can use/adapt in some OS/2 include file... I guess endianness conversion is performed even on OS/2 ;) Hmmm, I think you should not use these macros in userspace anyway! They are not portable an I'm pretty sure building SANE on any platform but Linux will fail. my 2 cents Gerhard
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Mon, 4 Dec 2006 19:28:43 +0100 (MET) Gerhard Jaeger gerh...@gjaeger.de wrote: __le16_to_cpup(), __u16,, __le32_to_cpup(), __u32, __le32 or __cpu_to_le32() are defined on OS/2. What do you suggest? mmm.. you should try to find similar macros we can use/adapt in some OS/2 include file... I guess endianness conversion is performed even on OS/2 ;) Hmmm, I think you should not use these macros in userspace anyway! They are not portable an I'm pretty sure building SANE on any platform but Linux will fail. mm. right. So I need to find some portable macros or include appropriate one in sane. any suggestion? -- Best regards, Alessandro Zummo, Tower Technologies - Turin, Italy http://www.towertech.it
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
Hi, two minor problems on OS/2 with current CVS (since 1. 12. 2006) 1. configure (generated by Autoconf 2.59e) fails and claims not to find grep (which is bogus) Running autoconf (2.59) locally solves the problem. 2 I have to add #include sys/socket.h to .sanei\sanei_tcp.c because otherwise AF_INET is not defined. 3. : *** No rule to make target `libsane-epson2.la', needed by `all'. Stop. probably a temporary problem until epson2 - code is added.. Thanks for listening Franz
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
Hi Alessandro, On Sat, 2 Dec 2006, Alessandro Zummo wrote: 1. configure (generated by Autoconf 2.59e) fails and claims not to find grep (which is bogus) Running autoconf (2.59) locally solves the problem. [...] dunno how to resolve the autoconf problem though... maybe only configure.in should be shipped. I have also fallen for this and been the Bad Man to break the OS/2 build by simply using the default autoconf installed on my debian box! The solution is just to install autoconf 2.59 (eg: from tarball), use that to generate ./configure, test and check in the result. There used to be a note about correct autoconf versions on the SANE website although a quick search just now failed to find it. cheers, Jon == Jon Chambers = http://www.jon.demon.co.uk, 020 8575 7097, 07931 961669 =
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Sat, 2 Dec 2006 14:23:05 +0100, Alessandro Zummo wrote: probably a temporary problem until epson2 - code is added.. ach! :( will commit a patch soon, thanks! Thanks, this leads to other problems ;-) epson2.c:53:27: asm/byteorder.h: No such file or directory epson2.c:841: warning: implicit declaration of function `__le16_to_cpup' epson2.c:841: error: `__u16' undeclared (first use in this function) epson2.c:1256: warning: implicit declaration of function `__le32_to_cpup' epson2.c:1256: error: `__u32' undeclared (first use in this function) epson2.c:4336: error: `__le32' undeclared (first use in this function) epson2.c:4338: warning: implicit declaration of function `__cpu_to_le32' (the warnings will change to errors about unresolved symbols, if I expand '__u16' to 'unsingned short' and so on. Franz
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Sat, 02 Dec 2006 20:04:37 +0100 (CET) Franz Bakan fba...@gmx.net wrote: On Sat, 2 Dec 2006 14:23:05 +0100, Alessandro Zummo wrote: probably a temporary problem until epson2 - code is added.. ach! :( will commit a patch soon, thanks! Thanks, this leads to other problems ;-) epson2.c:53:27: asm/byteorder.h: No such file or directory epson2.c:841: warning: implicit declaration of function `__le16_to_cpup' epson2.c:841: error: `__u16' undeclared (first use in this function) epson2.c:1256: warning: implicit declaration of function `__le32_to_cpup' epson2.c:1256: error: `__u32' undeclared (first use in this function) epson2.c:4336: error: `__le32' undeclared (first use in this function) epson2.c:4338: warning: implicit declaration of function `__cpu_to_le32' (the warnings will change to errors about unresolved symbols, if I expand '__u16' to 'unsingned short' and so on. well, you need to check where are those functions/types defined on os/2 :) -- Best regards, Alessandro Zummo, Tower Technologies - Turin, Italy http://www.towertech.it
[sane-devel] 3 minor problems with current (20061202) CVS code on OS/2
On Sat, 2 Dec 2006 20:31:06 +0100, Alessandro Zummo wrote: well, you need to check where are those functions/types defined on os/2 :) Yes, I know and already checked, but unfortunattely there is not place where __le16_to_cpup(), __u16,, __le32_to_cpup(), __u32, __le32 or __cpu_to_le32() are defined on OS/2. What do you suggest? Franz