On 10/06/2016 10:57 AM, Amos Jeffries wrote: > Please add a check to the unit test testSBuf::testAppendSBuf() > to guarantee that the (*this = S) assignment code path updates the store > reference count rather than doing a bit-wise copy of the SBuf.
I support that addition but do not have the time to implement it right now. It would also be nice to add a test case that the optimized assignment path does _not_ happen for pre-allocated SBufs. >> P.S. AFAICT, there is no guarantee that SBuf using GetStorePrototype() >> isEmpty so checking both conditions seems necessary to me. > That sounds like a bug to me. If the initial prototype is anything but > empty then the resulting new/clear'ed SBuf might end up with some random > data prefix. Followup patch? Not a bug AFAICT: The initial prototype storage might get dirty, but the freshly initialized SBuf will be _empty_, making any storage bytes irrelevant. Thank you, Alex. _______________________________________________ squid-dev mailing list squid-dev@lists.squid-cache.org http://lists.squid-cache.org/listinfo/squid-dev