04:13
<
gamersg >
jonasschnelli: u there?
04:15
<
gamersg >
sipa: u there?
04:16
<
gamersg >
is it ok if i do a getCommands(std::vector& list)
04:16
<
gamersg >
i prefer it to returning a vector, more efficient
04:17
<
sipa >
we'll switch to C++11 soon, in which you can use move semantics for that
04:17
<
sipa >
so i don't really care
04:17
<
gamersg >
move semantics will apply automatically i believe, so the efficiency gain would be automatic, amirite
04:18
<
sipa >
in cases where it applies automatically, i think copy elision would already apply right now as well
04:21
<
gamersg >
sipa: copy ellision?
04:21
<
gamersg >
u mean compiler optimisation?
04:22
<
sipa >
if you have a function f() { return some_vector; } ... vector v = f();
04:23
<
sipa >
then C++ semantics specify that 2 copies occur (once to copy some_vector to the return value, and one to assign the return value to v
04:23
<
sipa >
however, the compiler is allowed to not issue those copy operations and bypass the assignments
04:23
<
sipa >
which in practice always happens
04:23
<
sipa >
i dislike it because it's non obvious where that applies
04:23
<
sipa >
move semantics are cleaner
04:25
<
gamersg >
ic, alrite then ill stick to returning a vector
05:56
<
GitHub57 >
bitcoin/master 086da92 Suhas Daftuar: Add tags to mempool's mapTx indices
05:56
<
GitHub57 >
bitcoin/master 9f33dba Pieter Wuille: Merge #7539: Add tags to mempool's mapTx indices...