On Apr 25, 2009, at 1:53 PM, Richard Haselgrove wrote: >> I know, Richard you don't see as much need for this, but, this is >> the reason I cannot do more on the other end. >> >> Oh, and I won't even mention the waste of compute time doing >> nothing once every 10 seconds ... > > OK, I'm convinced - especially if there's a guard time using that > interesting 'now' that doesn't work either.
I know I have a prejudice, but, it has been my observation the C programmers tend to "sloppy" programming because the language "allows" it. The problem is that the lack of rigor leads to bugs. Is "now" really the time variable/function, or is it the value zero ... were it me, I would want the explicit and correct function just to be sure. I used to do explicit type-coercion and always were being asked why I did that when the compiler would do it for me ... of course the point was it reminded me what *I* really thought I was doing ... so if I went back to the code I knew what I really thought I was doing ... I hate assumed "magic" ... Actually it is a good thing that the thing does not work ... it would prevent rescheduling on task end for 60 seconds ... :) Not only is it broken, but it is wrong ... > The reschedule code is called too often: that needs debugging too. > But please don't stop at that point: the task switch timeslice is > still with us (i,e, you still have to earn that lunch.....lol!) I know, and I agree ... but I cannot find things in the logs because there is simply too much trash in there I am looking for a needle in the haystack. I cannot separate the bad calls from the simply repetitive wasted time calls. > If there was an easy way to get hold of a build environment, I'd be > tempted to put in some more debug output for a while, both to output > the raw values of the variables involved, and to validate whether > any of the alternative exit paths are ever followed. Before my last major collapse I was planning to tackle that part of documentation ... sadly it will likely never be done now as I am no longer up for hard things ... I can only do an hour or so a day on good days ... _______________________________________________ boinc_dev mailing list [email protected] http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
