on Wed May 02 2007, Rene Rivera
Boost Community,
The Boost Conference will be here in very short order! And I need to get feedback for the "Testing Boost" sprint we'll be doing http://boostcon.com/program/sessions#rivera-testing-sprint. If you have any ideas, concerns, off the wall comments, etc. at *minimum* send them directly to me at "rrivera/acm.org". Why that account? Well because that will be the one account I can guarantee I will access to during the conference, and hence people can send me stuff up to the last minute. Not that I'm encouraging lateness, but I know how busy we all are ;-)
If you are going to send it to the Boost dev list you will also want to eventually send me a summary follow up, if needed, from the ensuing discussions.
Okay, here are some issues I think ought to be solved, in no particular order, some (much) more important than others: * inaccurate header dependency tracing impedes useful incremental testing * report usability is poor for most constituencies. This point could be explored in much more detail. * no mechanism to pinpoint the checkin that caused a regression * no mechanism for communicating that tests are known not to work on some platform (toolset+os) to the build system, and so avoid building them * generating XML by parsing the jam log is fragile and prevents the use of multiple build processes (-jN). This one should be almost embarassingly easy to fix. * The smallest granularity of test that any tester can contribute is the entire Boost suite for one toolset, which makes for longer turnaround. * No way for a developer to request testing results for a specific branch * No "nanny system" that bothers the developer who causes a regression until it is fixed. I think I've thought of others from time-to-time; these are just the ones off the top of my head this morning. -- Dave Abrahams Boost Consulting www.boost-consulting.com Don't Miss BoostCon 2007! ==> http://www.boostcon.com