On Wed, Nov 20, 2013 at 7:32 PM, Daniel Jasper <[email protected]> wrote:
> > > > On Wed, Nov 20, 2013 at 9:28 AM, Alexander Kornienko <[email protected]>wrote: > >> On Wed, Nov 20, 2013 at 6:12 PM, Daniel Jasper <[email protected]>wrote: >> >>> >>>> @@ -681,6 +681,7 @@ void UnwrappedLineParser::parseStructura >>>> Style.BreakBeforeBraces == FormatStyle::BS_Stroustrup || >>>> Style.BreakBeforeBraces == FormatStyle::BS_Allman) >>>> addUnwrappedLine(); >>>> >>> >>> Does it still make sense to report the "{" as its own unwrapped line? >>> Seems a bit convoluted to first report multiple lines and then merge them >>> afterwards. I think this would make the merging code simpler. >>> >> >> It also seemed strange to me. Should we instead handle BreakBeforeBraces >> in TokenAnnotator? This will require adding TokenType values for braces >> starting namespaces, classes/structs and, probably, enums. I can play with >> this a bit, it you think it makes sense. >> > > I might have already done this for enums. I don't think it is essential to > add token types for all of these as e.g. enums and namespaces are really > easy to detect. But adding token types might be the cleaner solution. I > think that this makes sense but I remember having some kind of debate over > this with Manuel, so he might have an opinion. > I think the main motivation was that we didn't need it at the point, and we didn't have all the nuts and bolts in place to transfer the information precisely enough to the layouter... Now seems to be a good time to change this... :)
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
