[Lldb-commits] [PATCH] Fix warning about the use of mktemp and make platform agnostic by adding and using PipeBase::CreateWithUniqueName.

[Zachary Turner]

On Wed Feb 04 2015 at 10:39:58 AM Robert Flack <flackr at gmail.com> wrote:

> In http://reviews.llvm.org/D7348#118521, @zturner wrote:
>
> > Hi Robert, by any chance did you actually try compiling this on Windows?
> >  The reason I ask is that it just occurred to me that this will now
> require
> >  linking in rpcrt4.lib, which I don't think happens by default.  So this
> >  will cause linker errors.  If you tested it and it compiled, then no
> >  problem.  But if you didn't test it, then I think it's better to just go
> >  back to the way it was before, and I will do the UuidToString stuff
> later.
>
>
> I did, but it didn't get as far as the linking phase due to an unrelated
> error:
>
> D:\src\ll\llvm\tools\lldb\source\Target\ProcessLaunchInfo.cpp(347) :
> error C2065: 'O_CLOEXEC' : undeclared identifier
> [125/270] Building CXX object tools\ll...iles\lldbTarget.
> dir\ThreadPlan.cpp.obj
> ninja: build stopped: subcommand failed.
>
> So I'm not sure if there will be linker errors. I can go back to the way
> it was before on this.
>
Sounds good. I'll look into fixing this error.  I'm syncing now to see what
the status is of compiling at ToT.  I like your idea of not storing the
name, and just returning it as an output parameter.  Not everyone will care
about the name, so for them this doesn't impose a penalty on them.  Feel
free to change it to that way instead.  The usual paradigm for this is to
have the method take an llvm::SmallVector<char> & as an argument.  clear()
it then just write the value into it.  Callers will allocate a
SmallString<N> and pass it in by reference.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/lldb-commits/attachments/20150204/c23420fe/attachment.html>