> On June 15, 2015, 10:17 a.m., Bernd Mathiske wrote: > > 3rdparty/libprocess/3rdparty/stout/include/stout/flags/flags.hpp, line 146 > > <https://reviews.apache.org/r/34943/diff/4/?file=984516#file984516line146> > > > > Could we use some more concrete typing like this here? > > > > std::function<Option<Error>(Option<T>)>&& validate > > > > Similar types could then also be used in all other places where "F&&" > > appears below.
I can't use a well defined type here because the compiler can't deduce the 'T' inside Option<T> all by itself when using lambdas, instead the compiler needs the caller to explicitly specify the parameter type, which unfortunately makes all the call sites extremely more verbose (e.g., 'add<std::string>(&name, "name", "Description.", "default");'). I'm going to do another pass here and add documentation during which I'll clearly capture the expected type of F. > On June 15, 2015, 10:17 a.m., Bernd Mathiske wrote: > > 3rdparty/libprocess/3rdparty/stout/include/stout/flags/flags.hpp, line 293 > > <https://reviews.apache.org/r/34943/diff/4/?file=984516#file984516line293> > > > > What if T1 is Path? Then fetch() only parses. BTW: what does this mean > > then? This does not seem to be explained in fetch.hpp or anywhere else. I updated the comment here. And yes, we need more documentation around 'fetch' and 'parse' too. On it's way! > On June 15, 2015, 10:17 a.m., Bernd Mathiske wrote: > > 3rdparty/libprocess/3rdparty/stout/include/stout/flags/flags.hpp, line 286 > > <https://reviews.apache.org/r/34943/diff/4/?file=984516#file984516line286> > > > > I wonder if we need this comment at all. What it says seems rather > > obvious. You cannot assign a function value that does not match the given > > signature. > > > > It seems that this comment is a remnant, because a less trivial variant > > of it follows further down below. I cleaned this up, see my coments to Niklas below. - Benjamin ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/34943/#review87907 ----------------------------------------------------------- On June 15, 2015, 5:52 p.m., Benjamin Hindman wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/34943/ > ----------------------------------------------------------- > > (Updated June 15, 2015, 5:52 p.m.) > > > Review request for mesos and Niklas Nielsen. > > > Repository: mesos > > > Description > ------- > > Also refactored existing 'lambda::bind' arguments to use C++11 lambdas, > enabling us to get rid of our "loader" and "stringifier" functors. > > > Diffs > ----- > > 3rdparty/libprocess/3rdparty/stout/include/Makefile.am > 6ac2f04a6a1d8db1725972a3b4b46a0dd734566f > 3rdparty/libprocess/3rdparty/stout/include/stout/flags/flag.hpp > 87606d860dea3235ec70d127d3379065d42dc90b > 3rdparty/libprocess/3rdparty/stout/include/stout/flags/flags.hpp > ee855da6128c2d671eb2fc7eaa2c0aab2341aebb > 3rdparty/libprocess/3rdparty/stout/include/stout/flags/loader.hpp > 51d3ab020bd2bae1f12b66cac0da5383c813d5d7 > 3rdparty/libprocess/3rdparty/stout/include/stout/flags/stringifier.hpp > fda5ae1328a23a4371475652f921341dce7448d5 > 3rdparty/libprocess/3rdparty/stout/tests/flags_tests.cpp > 80450185f60c5b273face490e0bb9e695b0cb984 > > Diff: https://reviews.apache.org/r/34943/diff/ > > > Testing > ------- > > make check > > > Thanks, > > Benjamin Hindman > >
