Splint

Splint

[alin.s]

Have somebody tried to make an exhaustive test of splint in all emacs C
source code?

Does it worth inserting such scripts in Makefile.am ?




Alin


-- 
View this message in context: http://old.nabble.com/Splint-tp27057818p27057818.html
Sent from the Emacs - Dev mailing list archive at Nabble.com.

Re: Splint

[Ken Raeburn]

On Jan 7, 2010, at 05:43, alin.s wrote:
> Have somebody tried to make an exhaustive test of splint in all  
> emacs C
> source code?
>
> Does it worth inserting such scripts in Makefile.am ?

I played with splint a few years back (not for Emacs but another  
project).  It's an interesting tool, but the lack of maintenance and  
some issues in how it works were a bit of a problem.  For example, I  
found lots of annotations in the code were needed, and often  
declarations needed to be broken up with typedefs so you could  
annotate a pointer-to-pointer type properly.  I'd suggest looking for  
a better tool, unless there's been seriously renewed efforts to  
improve splint.

While Coverity's stuff is commercial, they've offered its use to some  
open-source/free software projects.  In fact, as I recall, Emacs was  
done some time ago, and a bunch of problems brought to this list,  
though I don't know if anyone is monitoring to see how we fare now.

Ken

Re: Splint

[Dan Nicolaescu]

Ken Raeburn <raeburn@raeburn.org> writes:

  > While Coverity's stuff is commercial, they've offered its use to some
  > open-source/free software projects.  In fact, as I recall, Emacs was
  > done some time ago, and a bunch of problems brought to this list,
  > though I don't know if anyone is monitoring to see how we fare now.

Unfortunately the last run was in Oct 2008.  I've sent a few emails to
the admin list asking to restart running.  No reply...

Re: Splint

[Richard Stallman]

    While Coverity's stuff is commercial,

Do you mean "proprietary"?  They are not the same.  There is nothing
wrong with a program's being commercial.  See
http://www.gnu.org/philosophy/words-to-avoid.html.

					  they've offered its use to some  
    open-source/free software projects.

See http://www.gnu.org//philosophy/mcvoy.html
for the story about another similar offer,
and the harm that was done when someone else accepted it.

Re: Splint

[Ken Raeburn]

On Jan 8, 2010, at 08:17, Richard Stallman wrote:
>    While Coverity's stuff is commercial,
>
> Do you mean "proprietary"?  They are not the same.  There is nothing
> wrong with a program's being commercial.  See
> http://www.gnu.org/philosophy/words-to-avoid.html.

Yes, sorry, I mis-spoke.

> 					  they've offered its use to some
>    open-source/free software projects.
>
> See http://www.gnu.org//philosophy/mcvoy.html
> for the story about another similar offer,
> and the harm that was done when someone else accepted it.

Yep, there is that issue.  At least in this case it is still practical  
and easy for someone to use splint and other free or at least open- 
source software, while someone else uses something like Coverity's  
tools.  Unlike Bitkeeper and Linux kernel source access, the  
proprietary tools shouldn't intrude significantly on those who don't  
want to use it in their development work.

I'd still love to see a free tool to do this work.  But when I did a  
survey for work a few years back, my impression was that the free  
tools were all poor; splint appeared to be one of the best of the lot,  
and was quite a pain to try to use.  Perhaps with recent work on gcc,  
llvm and other tools, the situation can be changed.

Ken