Re: Stripping Kannel
If there is a need for a lighter weight binary with reduced driver support shouldn't this be something put into the build process so that on compilation people can define the support and drivers they want to build in their binaries. It would be a painful process the first time it is done but with #ifdef it could be done quite nicely - and this would avoid the split builds that have hit projects like Mame so painfully. Alex Hi, I made some progress in stripping kannel. I have removed some drivers for smsc protocol, I got bearerbox binary of reduced size (around 170 KB less). The changes made are
Re: Stripping Kannel
Alex Judd wrote: If there is a need for a lighter weight binary with reduced driver support shouldn't this be something put into the build process so that on compilation people can define the support and drivers they want to build in their binaries. Yes, definitly. This would add a pro to the re-structuring of the build process using GNU autotools (autoconf, automake, etc.) with according sub-directories for the single SMSC modules. It's a bit picky to re-arrange it now, since it got complex enough. But I'm anyway trying to port the current Kannel source code tree to be GNU automake compatible. Are there any developers who want to join on that development? I guess there has to be someone who did extensive GNU autotool development already? Stipe [EMAIL PROTECTED] --- Wapme Systems AG Münsterstr. 248 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Stripping Kannel
Hi, I made some progress in stripping kannel. I have removed some drivers for smsc protocol, I got bearerbox binary of reduced size (around 170 KB less). The changes made are 1. smsc_p.h Removed entry for CIMD, CIMD2, Sema 4.5 Sema OIS 5.0 2. smsc.c Removed entry for CIMD, CIMD2, Sema 4.5 Sema OIS 5.0 3. smsc_cimd.c (deleted) 4. smsc_cimd2.c (deleted) 5. smsc_sema.h (deleted) 6. smsc_sema.c (deleted) 7. smsc_ois.c (deleted) Similarly I'll do for EMI EMI_IP. But what about SMPP and HTTP, as smsc.c does not have any entry for them (smsc-type). Can any one guide me thru this, as I want only AT drivers for my current work. Later I would try to remove WAP stack part from bearerbox. Thanks in advance. Shridhar Raju
Re: Stripping Kannel
so you would agree to have a configure switch that strips down everything WAP related in bearerbox, right?! I also agree with that! here we only use kannel for SMS purposes (for now ;) ) and it will be great if we don't need to compile wap related features! cheers, Valter Santos - Original Message - From: Stipe Tolj [EMAIL PROTECTED] Cc: Kannel Developers [EMAIL PROTECTED] Sent: Friday, January 04, 2002 2:15 PM Subject: Re: Stripping Kannel Angel Fradejas wrote: I agree this is an interesting thing. We also use kannel as a SMS gateway only, and striping down everything wap related makes sense to us, too. so you would agree to have a configure switch that strips down everything WAP related in bearerbox, right?! Nick, who is currently in charge of updating the wishlist or the TODO file? I'll put a item on my Kannel's new web site features for this. Stipe [EMAIL PROTECTED] --- Wapme Systems AG Münsterstr. 248 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
RE: Stripping Kannel
Stipe Tolj wrote: I agree this is an interesting thing. We also use kannel as a SMS gateway only, and striping down everything wap related makes sense to us, too. so you would agree to have a configure switch that strips down everything WAP related in bearerbox, right?! Yes, that would be fine. Regarding to the exclusion of some smsc drivers from bearerbox, I agree with you, I don't see any advantage in a --with-smsc-only=[xxx] flag. Angel.
Re: Stripping Kannel
Hi Stipe, As suggested by u, I defined #define KANNEL_NO_WAP 1 in config.h file, which was generated by configure script. But result was same, make compiled everything in gw/ and made all executables (bearerbox, wapbox smsbox). Little reading into bearerbox.c, we found that there r mention of KANNEL_NO_WAP KANNEL_NO_SMS. Any advice regarding this. Our requirement is to have only bearerbox and smsbox, and moreover bearerbox should not have WAP stack part into it. Similarly for smsbox, we want only select drivers like AT, CIMD2. Any comment from the group will definitely help us find our path. Shridhar Raju
Stripping Kannel
Hi, We have been using Kannel 1.1.5 and cvs for SMS applications. It has been working fine with AT and CIMD2 driver. Is it possible to strip Kannel to bare minimum like bearerbox will have only required SMSC driver and no WDP layer. If yes, how to do it. I thought about some options like 1. Changes in Makefile 2. Commenting whole driver files before compiling 3. And of course going thru whole source code. My ideas must be vague, but there has to be some way to do it. Please comment and guide me. Shridhar Raju
Re: Stripping Kannel
We have been using Kannel 1.1.5 and cvs for SMS applications. It has been working fine with AT and CIMD2 driver. Is it possible to strip Kannel to bare minimum like bearerbox will have only required SMSC driver and no WDP layer. If yes, how to do it. I thought about some options like 1. Changes in Makefile 2. Commenting whole driver files before compiling 3. And of course going thru whole source code. My ideas must be vague, but there has to be some way to do it. Please comment and guide me. yes there is a simple way, at least to strip down bearerbox to use only wap or sms features. bearerbox.c has a couple of conditional #define checkings for that. To use them specifiy manually within config.h: #define KANNEL_NO_WAP 1 to exclude compilation of WAP specific code within bearerbox.c and #define KANNE_NO_SMS 1 to exclude the SMS specific code. This has not been intended for what you actualy want, it was for debugging and testing, but should do at least a strip down. We may extend this idea for bearerbox to strip down to only needed function, even if I don't think that it is necessary for some realy mattering concerns. We may add a --with-bearerbox-mode=[sms|wap] configure switch which sets a #define in config.h accordingly and those may be tested for the conditional compile for bearerbox.c. Any comments on that from the list? Has anyone real application demands on this? Stipe [EMAIL PROTECTED] --- Wapme Systems AG Münsterstr. 248 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are
Re: Stripping Kannel
Hi Stipe, Thanks for the reply, it was very encourging for us to start doing something like this. It would be better idea, I guess, having --with-bearerbox-mode=[sms|wap] for the configuration. Any more suggestion from the list would be appreciated regarding striping of Kannel, which I have wrote in my last mail. Shridhar Raju We have been using Kannel 1.1.5 and cvs for SMS applications. It has been working fine with AT and CIMD2 driver. Is it possible to strip Kannel to bare minimum like bearerbox will have only required SMSC driver and no WDP layer. If yes, how to do it. I thought about some options like 1. Changes in Makefile 2. Commenting whole driver files before compiling 3. And of course going thru whole source code. My ideas must be vague, but there has to be some way to do it. Please comment and guide me. yes there is a simple way, at least to strip down bearerbox to use only wap or sms features. bearerbox.c has a couple of conditional #define checkings for that. To use them specifiy manually within config.h: #define KANNEL_NO_WAP 1 to exclude compilation of WAP specific code within bearerbox.c and #define KANNE_NO_SMS 1 to exclude the SMS specific code. This has not been intended for what you actualy want, it was for debugging and testing, but should do at least a strip down. We may extend this idea for bearerbox to strip down to only needed function, even if I don't think that it is necessary for some realy mattering concerns. We may add a --with-bearerbox-mode=[sms|wap] configure switch which sets a #define in config.h accordingly and those may be tested for the conditional compile for bearerbox.c. Any comments on that from the list? Has anyone real application demands on this? Stipe [EMAIL PROTECTED] --- Wapme Systems AG Münsterstr. 248 40470 Düsseldorf Tel: +49-211-74845-0 Fax: +49-211-74845-299 E-Mail: [EMAIL PROTECTED] Internet: http://www.wapme-systems.de --- wapme.net - wherever you are