[cfe-commits] r72146 - in /cfe/trunk/docs: InternalsManual.html PCHInternals.html

Ted Kremenek kremenek at apple.com
Wed May 20 03:50:34 PDT 2009


On May 19, 2009, at 5:19 PM, Eli Friedman wrote:

> On Tue, May 19, 2009 at 5:16 PM, Douglas Gregor <dgregor at apple.com>  
> wrote:
>> -<p>Clang's current implementation of precompiled headers, known as  
>> PTH, is
>> -documented <a href="PTHInternals.html">here</a>.</p>
>> +<p>Clang supports two implementations of precompiled headers. The
>> +   default implementation, precompiled headers (<a
>> +    href="PCHInternals.html">PCH</a>) uses a serialized  
>> representation
>> +   of Clang's internal data structures, encoded with the <a
>> +    href="http://llvm.org/docs/BitCodeFormat.html">LLVM bitstream
>> +   format</a>. Pretokenized headers (<a
>> +    href="PTHInternals.html">PTH</a>), on the other hand, contain a
>> +   serialized representation of the tokens encountered when
>> +   preprocessing a header (and anything that header includes).</p>
>
> Long-term, are we planning to keep around PTH, or discard it as a
> failed experiment?  Or has it not been discussed yet?

The offline discussion that we have had so far is that there is  
potential value in their coexistence.  For example, PTH can be used to  
cache the headers for an entire system in an architecture-independent.  
language-independent way (i.e., same PTH headers for C, Objective-C,  
and C++), which can be leveraged when generating PCH files specific to  
a given project.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20090520/c8481c68/attachment.html>


More information about the cfe-commits mailing list