The patch seems to be working. Thanks! #276 can be closed. On Wed, Sep 16, 2009 at 7:37 AM, David Anderson <[email protected]>wrote:
> I checked in a (possible) fix: [19055]. > Rytis, please let me know if this works. > -- David > > > Rattledagger wrote: > >> On Mon, 14 Sep 2009 20:17:21 +0300, you wrote: >> >> The project I'm working on requires that exactly one result is sent for >>> each >>> workunit. So I set max_total_results=1, max_success_results=1, >>> max_error_results=0. Workunit dies if a result is an error, goes as valid >>> if >>> a result runs fine, however, there is one case where this does not work: >>> timed out results. In case a result times out, a new result is created, >>> breaking max_total_results rule. >>> Can anything be done about that? >>> >> >> Well, atleast in my opinion, this should in theory be a fairly easy bug to >> fix... >> >> ... but, seeing it was reported back in 2004, and still no-one has fixed >> this >> bug, in practice it seems to be a really really hard bug to fix... >> >> For a bug-report, and my attempt on a fix (not correct C, but the logic >> behind >> it should likely still be understandable...), see >> http://boinc.berkeley.edu/trac/ticket/276 >> >> Hmm, quick look, now the relevant file to change is transitioner.cpp and >> line >> 368 and onwards. >> >> >> > -- Rytis Slatkevičius +370 670 77777 ICQ 270359185; MSN [email protected]; GTalk [email protected]; Skype rytis.s (text only) _______________________________________________ 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.
