Re: CVS Feature Branch - Windows Build Warnings Growing 5-Oct-2004

[Derek Robert Price]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mark D. Baushke wrote:

> Conrad T. Pino <Conrad@Pino.com> writes:
>
> >h:\conrad\projects\cvs-1.12\lib\xstrdup.c(33) : warning C4013:
> 'xclone' undefined; assuming extern returning int
> >h:\conrad\projects\cvs-1.12\lib\xstrdup.c(33) : warning C4047:
> 'return' : 'char *' differs in levels of indirection from 'int '
>
>
> It looks like Derek needs to add a copy of xclone to the cvs system...


No, GNULIB moved the xstrdup() function into xmalloc.c and I didn't
notice.  The xclone() function has been replaced by xmemdup().  I've
now removed xstrdup.c from the repository, which should fix your
problem once you remove it from the build files.

I'm currently setting up an auto-update mechanism which should notice
when files are added and removed from the GNULIB modules we use.

> >yesno.c
> >Creating library...
> >xmalloc.obj : warning LNK4006: _xstrdup already defined in
> xstrdup.obj; second definition ignored


Likewise.

> >--------------------Configuration: libdiff - Win32
> Debug--------------------
> >Compiling...
> >analyze.c
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(430) : warning C4018:
> '<' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(432) : warning C4018:
> '<' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(469) : warning C4018:
> '>' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(497) : warning C4018:
> '<' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(506) : warning C4018:
> '>' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(517) : warning C4018:
> '>' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(537) : warning C4018:
> '<' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(550) : warning C4018:
> '<' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(568) : warning C4018:
> '<' : signed/unsigned mismatch
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(1060) : warning C4090:
> 'function' : different 'const' qualifiers
> >H:\Conrad\Projects\cvs-1.12\diff\analyze.c(1060) : warning C4022:
> 'free' : pointer mismatch for actual parameter 1
> >cmpbuf.c
> >context.c
> >diff.c
> >H:\Conrad\Projects\cvs-1.12\diff\diff.c(603) : warning C4013:
> 'error' undefined; assuming extern returning int
>
>
> Hmmm... this is bad... I would have expected error() to be included...


I think diff was using the error() CVS provided...  Did one of my
recent changes prevent it from being linked?  Perhaps one of the other
errors we are discussing cascaded and prevented it from being compiled
and this will go away as they are fixed?

> >exithandle.c
> >H:\Conrad\Projects\cvs-1.12\src\exithandle.c(47) : warning C4028:
> formal parameter 1 different from declaration
> >H:\Conrad\Projects\cvs-1.12\src\exithandle.c(47) : warning C4024:
> 'signals_register' : different types for formal and actual
> >parameter 1
>
>
> We probably need to do the 'right' thing here to make everyone happy nad
> not try to share the same functions for both the signals_register() and
> the atexit() functions.


Actually, I'm pretty sure two sets of functions are not needed.  All
the functions should only be registered with atexit() and
signals_register() should get a single function which only blocks
interrupts (I think this may be unnecessary from the signal handler -
interrupts are already blocked) then calls exit() (which will cause
the atexit()) functions to be called.  The atexit() functions will
probably have to be altered to block interrupts first thing as well
since the cleanup functions are not designed to be called twice.
Finally, I think there is a little cruft that was intended to prevent
multiple executions under the current setup which could be removed
from the handlers.

I've been meaning to do this for awhile but haven't found the time.

> ...
>
> >H:\Conrad\Projects\cvs-1.12\src\main.c(1207) : warning C4013:
> 'nstrftime' undefined; assuming extern returning int
>
>
> I am not sure why nstrftime is not providing a prototype...


I'm fairly certain this is a GNULIB problem if anyone would like to
double-check me and take the time to write a proper bug report.

Cheers,

Derek
- --
                *8^)

Email: derek@ximbiot.com

Get CVS support at <http://ximbiot.com>!
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFBY/xrLD1OTBfyMaQRAn4YAKCJBH3pFxHP63RhZXACzDkNtEcnAwCfXLLq
aaMcpFCNqJZWvQ9j3F4Dzd8=
=WykF
-----END PGP SIGNATURE-----