On Tue, Aug 13, 2013 at 3:34 PM, Ehsan Akhgari <ehsan.akhg...@gmail.com> wrote: > > Reducing our #include dependency will help speed up the compilation for > C/C++ files, and will also help us build fewer files after each change. We > have a tracker bug (785103) for improvements to the #include dependencies, > and we have data in bug 901132 to show you where to improve first (see > https://bug901132.bugzilla.mozilla.org/attachment.cgi?id=789006 and the > discussion around that.)
Specifically, https://bugzilla.mozilla.org/show_bug.cgi?id=901132#c11 explains what that file contains: "This is a list of all include relations, ranked by their "criticality": the number of source files that would no longer include this header if the #include directive were dropped." In other words, it ranks each #include statement according to how much compilation/recompilation would be saved if the #include statement was removed. Excellent! > Anybody who knows C++ can help with this To a degree, certainly. There are three categories of #includes: (a) Can be removed immediately. (b) Can be removed with a small amount of effort (e.g. moving some code around). (c) Cannot be removed, or not without major effort. You don't need any component expertise to determine those in category (a), but expertise can help a lot to identify (and remedy) those in category (b). Nick _______________________________________________ dev-platform mailing list dev-platform@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-platform
