Re: [Qemu-devel] [PATCH] trace: fix group name generation

[Greg Kurz <groug@kaod.org>]

On Sat, 15 Oct 2016 13:03:10 +0100
Peter Maydell <peter.maydell@linaro.org> wrote:

> On 14 October 2016 at 23:06, Greg Kurz <groug@kaod.org> wrote:
> > On Fri, 14 Oct 2016 16:31:01 -0500
> > Eric Blake <eblake@redhat.com> wrote:
> >  
> >> On 10/14/2016 04:26 PM, Greg Kurz wrote:  
> >> > Since commit "80dd5c4918ab trace: introduce a formal group name for trace
> >> > events", tracetool generates C variable names and macro definitions out
> >> > of the path to the trace-events-all file.  
> 
> >> > diff --git a/scripts/tracetool.py b/scripts/tracetool.py
> >> > index 629b2593c846..b81b834db924 100755
> >> > --- a/scripts/tracetool.py
> >> > +++ b/scripts/tracetool.py
> >> > @@ -70,7 +70,7 @@ def make_group_name(filename):
> >> >
> >> >      if dirname == "":
> >> >          return "common"
> >> > -    return re.sub(r"/|-", "_", dirname)
> >> > +    return "_" + re.sub(r"[^\w]", "_", dirname)  
> >>
> >> This STILL doesn't solve the complaint that the build is now dependent
> >> on the location.  Why can't we STRIP off any prefix prior to the in-tree
> >> portion of the naming that we know is sane, instead of munging the
> >> prefix but in the process creating source code that generates with
> >> different lengths?
> >>  
> >
> > Heh, because the complaint was about the build break :)
> >  
> >> Ideally, compiling twice, once in directory 'a', and the second time in
> >> directory 'aaaaaaaaaaaaaaaaaaaaaaaaaaaa', should not make a noticeable
> >> difference in the final size of the executable due to the difference in
> >> lengths of the debugging symbols used to record the longer name of the
> >> second directory being encoded into lots of macro names.
> >>  
> >
> > You're right. I'm okay to look for a way to fix the symbol variable size
> > issue, but I think this patch still makes sense as it makes tracetool
> > less fragile in case we add a directory with an illegal character to
> > the QEMU tree... and it fixes annoying build breaks (Paolo also hit
> > this and asked to revert 80dd5c4918ab in another mail).  
> 
> I agree with Eric that we should just not be putting
> the build directory name in these variables -- this
> looks like it's simply a bug to me, and we should fix it.
> 

As I was saying in the previous mail, I also agree with Eric :)

And it isn't even about the variable size of the debugging info,
but rather about the random names of the generated symbols...

> I think the chances of us adding a directory to the QEMU
> tree itself with a silly name are quite low (not least because
> if you do that then you get a handy build failure to tell
> you not to do that ;-))
> 

Fair enough, even if the error is a bit obscure at first sight... and
we already have a silly named directory (9pfs), which would be supported
with a leading '_'... but hopefully, it is not at the top level ;)

> thanks
> -- PMM

Cheers.

--
Greg