[Lldb-commits] [PATCH] D13066: DWARFASTParserClang::CompleteTypeFromDWARF: Handle incomplete baseclass or child

[Zachary Turner via lldb-commits]

Also,when you say the next iteration, is this something that is going to
happen at an unknown time in the future whenever you happen to touch it
again, or are you working on another iteration now?

On Thu, Sep 24, 2015 at 4:25 PM Zachary Turner <zturner at google.com> wrote:

> To be clear, I would like this Makefile to turn into the following:
>
> LEVEL = ../../../make
>
> CXX_SOURCES = main.cpp length.cpp
>
> DEBUG_INFO_FULL = True
> DEBUG_INFO_LIMITED = True
>
> And that's it.  You shouldn't need anything else.  Whatever needs to
> happen in Makefile.rules to make this work should be done.
>
> By default, DEBUG_INFO_FULL should be True and DEBUG_INFO_LIMITED should
> be false.  This will allow all other makefiles that don't care about the
> debug info to work as they normally do.  If neither variable is True, it
> won't use -g.  If only DEBUG_INFO_FULL is True, it will build one target
> and use -fno-limit-debug-info.  If only DEBUG_INFO_LIMITED is true, it will
> build one target and use -flimit-debug-info.  If both are true, it will
> build two targets, one for each case.
>
> Does this seem reasonable?
>
> On Thu, Sep 24, 2015 at 12:57 PM Zachary Turner <zturner at google.com>
> wrote:
>
>> zturner added a comment.
>>
>> I know, I've seen them in a few places myself.  But I think we should move
>> away from that, because it's a large source of portability problems
>>
>>
>> http://reviews.llvm.org/D13066
>>
>>
>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/lldb-commits/attachments/20150924/77e4e27a/attachment.html>