On Jan 30, 2014, at 11:41 AM, Phil Sorber <sor...@apache.org> wrote: > >> On Thu, Jan 30, 2014 at 11:37 AM, Igor Galić <i.ga...@brainsware.org> wrote: >> >> >> >> ----- Original Message ----- >>> https://code.google.com/p/cpp-btree/ >>> >>> I have been thinking about adding this to libts. It's licensed AL2 >> already. >> >> You mean to literally drop it in there, or require it as dependencies for >> people to build externally? > > Literally drop it in. It's not a library you can build. It's designed to > become part of your code base.
STL is generally ok, as long as it's not on the critical path, or bloats things unnecessarily . Any new or external library used on the CP should be carefully investigated for heap based memory allocations and lock contentions . I'd also make sure we avoid code / functionality duplication. If we bring in something new that's a superset of existing code, we should eliminate the old code. We've messed this up in the past, hence we still suffer with the TCL hash for no good reason. Cheers, -- Leif > > >> >>> It currently requires C++11 but there is also a patch to remove that >>> dependency that is AL2 licensed as well. >>> >>> I personally need this because it implements an ordered set where all our >>> hash table implementations are obviously unordered. Since std::map is off >>> the table, this seems like a good replacement. They even bench it against >>> std::map for comparison. >> >> Are you considering to replace the current uses of std::map too? > > Sure, if we have others. I have one use in a patch I am working on, but as > we have rejected the use of std::map outright, I plan to replace it before > I commit it. > > >> >>> Thoughts? >> >> -- >> Igor Galić >> >> Tel: +43 (0) 664 886 22 883 >> Mail: i.ga...@brainsware.org >> URL: http://brainsware.org/ >> GPG: 8716 7A9F 989B ABD5 100F 4008 F266 55D6 2998 1641 >> >>
