Re: [asterisk-dev] code-cleanup concerns
On Fri, 14 Apr 2006 22:28:24 -0400 Mark Sirota [EMAIL PROTECTED] wrote: --On Friday, April 14, 2006 4:28 PM -0700 Brian Degenhardt [EMAIL PROTECTED] wrote: Unless somebody can explain why for(;;) is inferior to do{}while(1), or why removing curly braces on one-line if statements is worth the trouble of patch maintainers everywhere, I think this just does more harm than good. Just FYI, from a 25-year C programming veteran: for (;;) is a more efficient loop than while (1) for some compilers, because there is no comparison to zero required. In modern reality most optimizers will make them identical, but it is conceivable that there might still be a system around that treats these literally. In any case, it's certainly clear to me that either of them is far more readable than do {} while (1)! Some compilers print a warning for while (1) as this is a constant expression. This warning is intended to catch typos like while (v = 1) instead of while (v == 1). The construct for (;;) does not generate a warning on any compiler I am aware of. Craig --- Craig Southeren Post Increment VoIP Consulting and Software [EMAIL PROTECTED] www.postincrement.com.au Phone: +61 243654666 ICQ: #86852844 Fax:+61 243673140 MSN: [EMAIL PROTECTED] Mobile: +61 417231046 It takes a man to suffer ignorance and smile. Be yourself, no matter what they say. Sting ___ --Bandwidth and Colocation provided by Easynews.com -- asterisk-dev mailing list To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-dev
Re: [Asterisk-Dev] DUNDi(tm) = Gnutella?
On Thu, 21 Oct 2004 08:23:55 -0400 Andrew Kohlsmith [EMAIL PROTECTED] wrote: On October 21, 2004 08:15 am, Craig Southeren wrote: Ummm...no, it's not even close to being basically the same as Gnutella. If anything, it's closer to DNS. Seems closer to Gnutella than DNS to me. DNS is hierarchial; DUNDi is not. Why do you say DUNDi isn't hierarchical? The notion of formal peering agreements forces a hierarchy (which may not be a simple tree like DNS, but will be a levelled hierarchy nonetheless), and caching of route data based on the TTL information makes it more efficient. DUNDi is not used for exchanging multi-megabyte media files - it's used for making very small routing requests. Gnutella's routing is also very small, but tons of routes adds up to big bandwidth, as Gnutella found out. Nodes in a DUNDi network will know whether or not they are authoritative for certain number sub-spaces, and hence deployers can make appropriate decisions for bandwidth, CPU etc at each node. Contrast this to the GnuTella network where random nodes could find themselves coopted as crucial nexii in the network without warning, with sometimes very negative results. Again, this is like DNS - links between DUNDi nodes are created as a result of explicit peering agreements, rather than node links being randomly created by automatic exchange of node addresses. Craig --- Craig Southeren [EMAIL PROTECTED] / [EMAIL PROTECTED] Phone: +61 243654666 ICQ: #86852844 Fax:+61 243673140 MSN: [EMAIL PROTECTED] Mobile: +61 417231046 Jabber: [EMAIL PROTECTED] Post Increment - Consulting Serviceshttp://www.postincrement.com Vox Gratia - The Open Source VoIP portal http://www.voxgratia.org Raving Of A Strange Mind - the VoIP blog http://www.southeren.com/blog ___ Asterisk-Dev mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-dev To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-dev
Re: [Asterisk-Dev] UK Caller ID patch and new CVS
As one of the core developers of the OpenH323 project, I agree with this view. Sometimes, forcibly integrating user contributed patches at any cost creates more problems than it's solves. It's not always a matter of increasing functionality at any cost - in the long term, an Open Source project is more about mantainability and ease of understanding than simply increasing functionality. Sometimes, it is better to wait until someone who *really* understands the code is able to create a patch that is fully integrated with the existing code, rather than just hacking in a change to solve a problem. I know that OpenH323 is a dirty word as far as some of the Asterisk developers are concerned, but this problem is common to both projects. Good patches are hard to find :) Craig On Fri, 23 Jul 2004 10:57:26 -0500 Dr. Rich Murphey [EMAIL PROTECTED] wrote: Perhaps it's not a question of elegant code but rather quality and maintainability. Once code goes in, those are real issues that take time and effort. I've seen technical integrity mistaken for dictatorship many times, and in the end the users suffer if it is replaced by political solutions. Believe me, you would be far worse off if such decisions were in the hands of a populist rather than an software architect. As an open source project one has every opportunity to change the code, but in submitting code, one must be able to take responsibility for long term issues. Please guys, focus on improving patches rather than asking the maintainers to lower standards. Regards, Rich --- Craig Southeren [EMAIL PROTECTED] / [EMAIL PROTECTED] Phone: +61 243654666 ICQ: #86852844 Fax:+61 243673140 MSN: [EMAIL PROTECTED] Mobile: +61 417231046 Jabber: [EMAIL PROTECTED] Post Increment - Consulting Serviceshttp://www.postincrement.com Vox Gratia - The Open Source VoIP portal http://www.voxgratia.org Raving Of A Strange Mind - the VoIP blog http://www.southeren.com/blog ___ Asterisk-Dev mailing list [EMAIL PROTECTED] http://lists.digium.com/mailman/listinfo/asterisk-dev To UNSUBSCRIBE or update options visit: http://lists.digium.com/mailman/listinfo/asterisk-dev