[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Autoscan fails cryptically on our code

From: Ralf Wildenhues
Subject: Re: Autoscan fails cryptically on our code
Date: Sat, 13 May 2006 19:49:04 +0200
User-agent: Mutt/1.5.11+cvs20060403

Hi Richard,

* Richard Brooksby wrote on Thu, May 11, 2006 at 07:14:19PM CEST:
> Following the instructions in the Goat Book at <http:// 
> sources.redhat.com/autobook/autobook/autobook_241.html#SEC241> I ran  
> "autoscan" on our source tree, which is visible at <http:// 
> www.ravenbrook.com/project/mps/master/code/>. 

FWIW, an URL to a tarball would have been more helpful for testing this.
I found this one,

but neither that nor the code in the directory you posted exposes your
issue for me.  I tested with CVS HEAD and 2.59.  Maybe you can post a
reduced test case that exposes this issue?

> >address@hidden autoscan -v
> >autoscan: srcdir = .
> >autoscan: cannot open <mpm.h: No such file or directory

> The file "mpm.h" is in the current directory.
> The angle bracket is suspicious.
> I think autoscan may be confused by the tagging conventions used for  
> cross-references within the sources.  We have many comments which  
> look like this:
> >* .source: See .design also.  PoolStruct and PoolClassStruct, the
> >* central types for this module, are defined in <code/mpmst.h>, the
> >* corresponding abstract types in <code/mpmtypes.h>.  Declarations and
> >* prototypes are in <code/mpm.h>.  Several functions have macro  
> >versions
> >* defined in <code/mpm.h>.  */

I think it should skip those comments.  autoscan is not perfect in
skipping comments, but should work reasonably.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]