lint tool for Ada?

lint tool for Ada?

[Jon Wilson, PRC]

Has anyone heard of a "lint-like" tool for Ada? (especially the free variety)

---
Jon R. Wilson                           Work: (205) 215-5477
Senior Computer Analyst                 Fax:  (205) 215-5511
PRC, Inc.                               jwilson@mtgy.gtegsc.com

Re: lint tool for Ada?

[Bill Yow]

In article <20nij1$esc@europa.eng.gtefsd.com>, jwilson@mtgy.gtegsc.com (Jon Wil
son, PRC) writes:
|> Has anyone heard of a "lint-like" tool for Ada? (especially the free variety
)
|> 

What would an Ada lint tool look for that is not already found by the 
compiler/run-time checks?

                                                Later,
                                                Bill Yow
                                                (713) 280-1591
                                                yow@pat.mdc.com
                                                byow@mcimail.com

My opinions are my own!

Re: lint tool for Ada?

[Jon Wilson, PRC]

In article 6390@aio.jsc.nasa.gov, yow@sweetpea.jsc.nasa.gov (Bill Yow) writes:
> In article <20nij1$esc@europa.eng.gtefsd.com>, jwilson@mtgy.gtegsc.com (Jon W
ilson, PRC) writes:
> |> Has anyone heard of a "lint-like" tool for Ada? (especially the free varie
ty)
> |> 
> 
> What would an Ada lint tool look for that is not already found by the 
> compiler/run-time checks?
> 

Wouldn't it be nice to find unreferenced variables, worthless withs, dead code,
 etc?
---
Jon R. Wilson                           Work: (205) 215-5477
Senior Computer Analyst                 Fax:  (205) 215-5511
PRC, Inc.                               jwilson@mtgy.gtegsc.com

Re: lint tool for Ada?

[Alex Blakemore]

In article <20nl8e$esc@europa.eng.gtefsd.com> jwilson@mtgy.gtegsc.com writes:
> Has anyone heard of a "lint-like" tool for Ada? (especially the free variety)
> Wouldn't it be nice to find unreferenced variables, worthless withs, dead cod
e, etc?
yes.

Marlestone Software Technology is finishing such a tool soon, currently
named the Ada Quality Toolset (there are other tools planned/in progress too)

Its been a research prototype - so price/availability hasnt been decided yet.

if you would like to be a beta tester or get more info, contact
Mark Dowson - dowson@marlstone.com or myself alex@cs.umd.edu

If you can make WADAS tomorrow afternoon, you can see a brief presentation.

The tools require ASIS and POSIX.

Disclaimer: I am not a disinterested party
-- 
Alex Blakemore       alex@cs.umd.edu        NeXT mail accepted
--------------------------------------------------------------
"Without an engaged and motivated human being at the keyboard,
the computer is just another dumb box."      William Raspberry

Re: lint tool for Ada?

[David Helken]

I would also like to know where a lint style tool could be found. Those who
underestimate its utility probably do not have to maintain Ada code. I have
often come accross with statements that were not needed - the compiler will not
tell you that some package is not referenced. I've seen type declarations that
have been superceded - the compiler will not tell you that a type is not used.
I've seen variables that are not referenced - the compiler does not catch
those either. The lint tools I've used also detected variables that were not
initialized before being referenced - the compiler could care? I'd like to see
such a tool that could also let me know that enumerated names were not
referenced in any 'image or 'value statements so that the space consumed in
some compiler for storing enumerated names could be freed thereby shrinking the
final code size. The binder usually will remove uncalled subprograms, but I
know of nothing that will eliminate unused variables. The point is, a lint tool
would be helpful for those of us who could use it. 

  Am I just using the wrong compiler? :-)

	David

Re: lint tool for Ada?

[cis.ohio-state.edu!news.sei.cmu.edu!ajpo.sei.cmu.edu!wellerd]

In article <C9GAvr.L05@dale.cts.com> helkenn@dale.cts.com (David Helken) writes
:
>The point is, a lint tool
>would be helpful for those of us who could use it. 
>

OK.  Adamat by Dynamics Research Corp will do "lint-like" checks for
you (and a lot more, for what it's worth).  Better be ready to open
your wallet though.

Regrettably, I don't have their phone number, but you might find it in
one of the info-ada files (anon ftp to ajpo.sei.cmu.edu).  It also
might (should?) be a FAQ.


-- 
-Comments above aren't neceessarily the opinion of the SEI, AJPO, or CAE-Link-
David Weller  |  Have you hugged your DRAGOON lately?
----I'm the Ultimate International Masochist: I speak Ada AND Esperanto!-----

Re: lint tool for Ada?

[agate!howland.reston.ans.net!darwin.sura.net!europa.eng.gtefsd.com!slc17!]

In article L05@dale.cts.com, helkenn@dale.cts.com (David Helken) writes:
> The point is, a lint tool
> would be helpful for those of us who could use it. 
> 
>   Am I just using the wrong compiler? :-)
> 

I think we both are. I was ignoring the people who didn't see the utility of a 
tool
like lint as either non-c programmers, or people who have a MUCH better compile
r
toolset than I, but based on the number of direct replies I got, there is quite
 a
market out there.

How about it compiler vendors? The only third-party tool that responded needed 
POSIX
and ASIS compliance. Either the compilers need to go this route so that third-p
arty
software can make Ada programmers lives less miserable, or provide toolsets tha
t
actually provided the functionality we need.

Before the flames... I know ONE compiler/toolset that meets these criteria. If 
you
are using it, you are quite lucky. I like their products, but I'm stuck with wh
at 
we have, just like a many others.


---
Jon R. Wilson                           Work: (205) 215-5477
Senior Computer Analyst                 Fax:  (205) 215-5511
PRC, Inc.                               jwilson@mtgy.gtegsc.com

Re: lint tool for Ada?

[vnet.IBM.COM]

>How about it compiler vendors? The only third-party tool that responded needed
 POSIX
>and ASIS compliance. Either the compilers need to go this route so that third-
party
>software can make Ada programmers lives less miserable, or provide toolsets th
at
>actually provided the functionality we need.

There was a panel discussion on tools like this at WADAS yesterday.  I
gather that for any meaningful analysis of the code ASIS is needed (or
it will have to be compiler specific) because you need to get at the
sublib to determine things like unused withs etc.  Ada Assured was
another interesting tool that does some of this kind of thing.  Mostly
based around a language sensitive editor.  It probably also requires
ASIS but I don't know about POSIX.

There were five guys on this panel each from a different company.
Ada Mat, Ada Assured, AQT (which is not a commercial product),
a company called Little Tree (don't remember the product name).
I forget the last.  Apparently there is a market out there.

********************************************************************
* These opinions are mine only.                     John Nestoriak *
********************************************************************

Re: lint tool for Ada?

[cis.ohio-state.edu!math.ohio-state.edu!sdd.hp.com!saimiri.primate.wisc.ed]

In article <19930701.105001.319@almaden.ibm.com> jnestoriak@vnet.IBM.COM writes
:
>>How about it compiler vendors? The only third-party tool that responded neede
d POSIX
>>and ASIS compliance. Either the compilers need to go this route so that third
-party
>>software can make Ada programmers lives less miserable, or provide toolsets t
hat
>>actually provided the functionality we need.
>
>There was a panel discussion on tools like this at WADAS yesterday.  I
>gather that for any meaningful analysis of the code ASIS is needed (or
>it will have to be compiler specific) because you need to get at the
>sublib .........

Or the tool (as the only two I've ever seen do) can parse the Ada source code 
itself.  Basically a compiler front-end with a report writer in place of a code
 generator.

Re: lint tool for Ada?

[Erik Magnuson]

In article <C9GAvr.L05@dale.cts.com> helkenn@dale.cts.com (David Helken) writes
:
>  Am I just using the wrong compiler? :-)
>
>	David

Maybe. Let's see now...

1. Unused objects
   - Rational will highlight all unused withs, declarations, and objects
     with a single key. 
   - Alsys cross reference utility will also flag unused objects
     including enumeration literals. (Although it flags as unused
     named numbers that are used only to define other named numbers.)

   I'm sure other vendors xref utilities do as well. In general, this
   feature requires the units to be compiled (possibly with debugging
   or some other extra options) and some way to specify
   which units to consider.

2. Uninitialized variables
   - ICC (w/optimization) will warn about potential uninitialized variables

3. Dead code
   - Alsys and ICC both warn about dead code (when they eliminate it)

I've traditionally included assessments of these features in my compiler
evaluations. This is from memory for the last several compilers I
worked with. I used to have a longer list but I'm sure it's out of
date by now. (If anyone else wants to jump in with what compilers do
what, I'd find it interesting.)

Should providing these features be the job of the Ada compiler vendor?

Pro:
  The compiler (and optimizer) often has to figure this out anyway.
  The compiler (or vendor) can take into account vendor specific stuff
  like pragmas.

Con:
  The Ada vendors have a hard enough time getting good compilers out
  the door! 
  Not all vendors will supply the same capabilities (IMHO support
  toolsets vary more than code quality.)

--
Erik