Again, thanks for root-causing this! Fixed the cause for the regression in r190126. However, I don't think we address this well in all cases. E.g. if getTopLevelRegion() had a parameter, we would still break after that call (and we "always" have). Please file a bug, if this comes up!
On Fri, Sep 6, 2013 at 1:20 AM, Tobias Grosser <[email protected]> wrote: > On 08/27/2013 07:24 AM, Daniel Jasper wrote: > >> Author: djasper >> Date: Tue Aug 27 09:24:43 2013 >> New Revision: 189353 >> >> URL: >> http://llvm.org/viewvc/llvm-**project?rev=189353&view=rev<http://llvm.org/viewvc/llvm-project?rev=189353&view=rev> >> Log: >> clang-format: Format segments of builder-type calls one per line. >> >> This fixes llvm.org/PR14818. >> >> Before: >> return llvm::StringSwitch<Reference::**Kind>(name) >> .StartsWith(".eh_frame_hdr", ORDER_EH_FRAMEHDR) >> .StartsWith(".eh_frame", ORDER_EH_FRAME) >> .StartsWith(".init", ORDER_INIT).StartsWith(".fini"**, >> ORDER_FINI) >> .StartsWith(".hash", ORDER_HASH).Default(ORDER_**TEXT); >> >> After: >> return llvm::StringSwitch<Reference::**Kind>(name) >> .StartsWith(".eh_frame_hdr", ORDER_EH_FRAMEHDR) >> .StartsWith(".eh_frame", ORDER_EH_FRAME) >> .StartsWith(".init", ORDER_INIT) >> .StartsWith(".fini", ORDER_FINI) >> .StartsWith(".hash", ORDER_HASH) >> .Default(ORDER_TEXT); >> > > This change introduced the following regression: > > Before: > > int foo() { > if (1) > if (1) > if (1) > O.indent(2 * (depth + 1)) > << "Node" > << static_cast<void *>(RI->getTopLevelRegion()->** > getBBNode(*BI)) > << ";\n"; > } > > After: > > int foo() { > if (1) > if (1) > if (1) > O.indent(2 * (depth + 1)) << "Node" > << static_cast<void > *>(RI->getTopLevelRegion() > > ->getBBNode(*BI)) > << ";\n"; > } > > I believe moving the << to a new line is better to read than adding a line > break before the call to ->getBBNode(). > > Cheers, > Tobias >
_______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
