> Having same message repeated tens or hundreds of >times per second because the server goes bananas it >is not useful to anyone, and can easily be automatically >suppressed. Handling this in the logging system means >it needs to be done only once instead of in every place >it can happen. If you want to make the thresholds >tunable it's also easier if it's handled in one place.
This is pretty general and something that everybody can pretty much agree to. I don't know how to translate this into specific code. From a technical perspective we're not switching to C++ and exceptions where something like this could be handled by a try/catch. In the particular case of polling failing(which is the *only* background activity in OpenOCD, this doesn't happen "everywhere"), it is also useful to see the first few error messages. So far I'm pretty confident that the backoff algorithm offers a *simple* localized solution to this generic problem. If in the future things develop, then that backoff algorithm can easily be tossed out and replaced by something new and more suitable. -- Øyvind Harboe US toll free 1-866-980-3434 / International +47 51 63 25 00 http://www.zylin.com/zy1000.html ARM7 ARM9 ARM11 XScale Cortex JTAG debugger and flash programmer _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
