[cfe-dev] libclang's Indexer assertion on translation unit in large C++ codebase

Daniel Jennings sekkusu at gmail.com
Thu Aug 25 16:06:35 PDT 2011


I'll try preprocessing the file before running it through the parser to see
if I can reproduce it.

Thanks for the help!


On Thu, Aug 25, 2011 at 4:04 PM, Douglas Gregor <dgregor at apple.com> wrote:

>
> On Aug 25, 2011, at 3:48 PM, Daniel Jennings wrote:
>
> > Hi all,
> >
> > I'm currently trying to run libclang's indexer over parts of an existing
> (large) codebase (that has never seen clang before) in order to run some
> metrics. I've got it working for some scenarios but in the process of
> getting it working I've run into a failed assertion in clang's AST code that
> I'm not sure how to track down. This problem only seems to exist when I try
> to run the indexer over parts of the real codebase; simple test CPP files
> work fine. As you'll see by the message, I'm using the python bindings for
> libclang, and I have my own locally built libclang.dll built under VS2010.
> >
> > Program: c:\Python27-32\python.exe
> > File: E:\llvm\tools\clang\include\clang/AST/DeclCXX.h
> > Line: 522
> > Expression: DefinitionData && "queried property of class with no
> definition"
> >
> > and
> >
> > libclang: crash detected during parsing: {
> >   'source_filename' : '(null)'
> >   'command_line_args' : ['u:\Code\Services\MyService\MyService.cpp'],
> >   'unsaved_files' : [],
> >   'options' : 0,
> > }
> >
> >
> > I'm not entirely sure how to track this down and figure out if I'm doing
> something wrong. I'm just trying to run the indexer over a single
> translation unit (a single CPP file) but that CPP file includes all of our
> normal core headers, etc. which I think may be the problem. By this point
> I'm very familiar with how the Python bindings work/wrap the libclang code,
> and I have my IDE set up for working with libclang code, so if you need me
> to poke around or add some sort of diagnostic code I can.
>
> This looks like Clang crashing during parsing. Can you reproduce the
> problem with preprocessed input? If so, please open a bug report so we can
> track down the issue.
>
>        - Doug
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20110825/d7209d5d/attachment.html>


More information about the cfe-dev mailing list