On Sun, Sep 8, 2013 at 4:29 PM, Nicholas Cameron <nick.r.came...@gmail.com> wrote: > > I don't think these kind of time improvements make it worth duplicating std > library code into mfbt, we may as well just pull in the headers and forget > about it. A caveat would be if it makes a significant difference on slower > systems. > > Given that improving what gets included via headers can make significant > difference to build time, this makes me wonder exactly what aspect of header > inclusion (if not size, which we should catch here) makes the difference.
My gut feeling is that the wins come from (a) cases where an additional include causes *huge* amounts of extra code to be pulled in, and (b) faster incremental builds due to fewer dependencies. (I've been focusing on (b) with the jsapi.h-dependence minimization.) In contrast, the inclusion of algorithm.h causes only a moderate amount of extra code to be pulled in. Nick _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
