[cfe-commits] [PATCH] Even More Abbreviations for PCH Serialization

[Douglas Gregor]

On Jun 3, 2011, at 3:33 PM, Jonathan Turner wrote:

> 
> On Jun 3, 2011, at 3:16 PM, Douglas Gregor wrote:
>>> 
>>> I'm being careful with any node types that use the TypeSourceInfo because the call to ASTWriter::AddTypeSourceInfo works differently if its first argument (which is D->getTypeSourceInfo() for us) is 0.   
>> 
>> Sure, but as long as it always puts *something* into the record, we should be fine.
> 
> Good point.  As long as we can assume that TypedefDecl will always have some TypeSourceInfo, we're fine.  I've added an assertion to note this.  If that's sounds good, I'll check in.

Why is TypedefDecl special in this regard? My assumption is that, if the TypeSourceInfo* is NULL, we'll just end up writing out one element in the VBR array that says "it's a NULL pointer". Is that not true?

	- Doug
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20110603/9fc4cdf5/attachment.html>