Lint

in the static analyzer market this is virtually free

PC-lint is good value, it can find a lot of potential bugs

my biggest issue with it is the low signal-to-noise (false positives) so only a few of us at work are willing to use it

Start with -w0 and work up to -w3 (don't use -w4)

Also you can tune the options file to get rid of any common false positives. It depends what compiler and config file you are using

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

All prospective users are not in your position. The $350 (or $1000 for Linux/Unix versions) is not practicable for retirees, curious teenagers, in fact most curious possibilities. This is not a criticism of Gimpel, because he is in business to make the most money, and it is his product.

And, as has been said before, to all practical purposes PC-LINT and Splint are the ONLY available static analyzers (apart from actual compilers). They are there to save the idiots, who are unable to maintain largely ISO compatible source.

--
 [mail]: Chuck F (cbfalconer at maineline dot net) 
 [page]: 
            Try the download section.

PC-Lint, at least, is capable of more than that.

Robert

And neither is any other static analyser. Neither is F1 or WRC or diamond collecting or horse racing. What is your point that some people can not afford the tools? This is the same for all activities.

No. Most curious people in the profession can afford PC-lint

Yes. The same with all tool suppliers.

This is completely un true. And apparently according to other posters Splint is a dead system

Then you do not understand static analysis

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

unfortunately it seems that you need to remove some desired rules to achieve this We do not put lint directives in the code! That is a big mistake in many ways.

Has this improved in the 9.0 version? We are still using the older version.

And there are plenty of other static code checkers, though most cost many times more than PC-Lint (depending on the project the extra cost can easily be justified).

Indeed it is a mistake in many ways to categorically refuse putting lint directives in the code.

C, as a language, isn't strong enough to express your code's intended behaviour quite as precisely as a tool like Lint would need it to know what to warn you about, and what not. So, short of inventing a technique for lint to read your mind, either you live with what you consider "spurious" warnings, or you _tell_ lint that you checked a particular warning in a particular piece of code, and that you're sure it does exactly what you want it to. That's what lint comments are about.

Exactly.

I don't understand CBF's problem that is not practicable for retirees, curious teenagers. SO WHAT Neither are the tools for most professions.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

On the other hand, an inexpensive, non-commercial license is one way to get a product into the hands of those who want to give it a turn around the block, in addition to the odd retiree or curious teenager. They don't even seem to offer as much as a limited-time/limited-linecount trial package; it's full price or nothing.

--
Rich Webb     Norfolk, VA

Don't forget as well that today's curious teenager is tomorrow's programmer with influence over which products to buy with the budget for a project or department.

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980's technology to a 21st century world

*LOL* I never met a programmer/engineer with enough influence to what (s)he wants.

And likely, if he got the tool as student for free, he sets up his mind, that it shall be always free _and_ will continue to use it, even if he is not student any more.

--
42Bastian
Do not email to bastian42@yahoo.com, it's a spam-only account :-)
Use @monlynx.de instead !

I never heard that students delete their student-versions of tools they bought/got.

But it seems it is not limited to a project. So 380USD is not really much even for a teenager (most of them pay more for their cellular phone bill in 2 months).

--
42Bastian
Do not email to bastian42@yahoo.com, it's a spam-only account :-)
Use @monlynx.de instead !

Although, when it comes to embedded systems, my interest is as a hobbyist (as my posting history will tell you), I am a normal commercial professional programmer and yes, there does seem to be a tendency to at least listen to new recruits, at least in the smaller outfits.

Whenever I encounter a junior person pushing a new idea or toolset, I for one, do firmly question them to make sure that they can justify what they are saying and are not just looking to add "used widget 2.x" to their CV, but they do get listened to - I for one am interested in listening to new ideas and seeing if there is anything in them.

Is there a more conservative attitude in embedded systems workplaces ?

It's been too long since I was a student to be able to accurately comment on this. :-)

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980's technology to a 21st century world

Very likely true. However, if the version license can't be upgraded (at least, not at the student price) and/or is node-locked then there's a reasonable likelihood of one (or more) future full-price sales.

Certainly MatLAB's dominance could be attributed, at least in part, to a smart plan to get copies into the hands of lots of future users.

For one data point, I've probably got nose hairs that qualify as being teenagers and my general philosophy is "He who dies with the most tools, wins." But, for all the wonderful marketing that Gimpel does, how well does the tool fit into my own peculiar workflow? How useful will it be

*to me*? I'd just as soon NOT buy the full-priced version (particularly since it does seem to be node-locked) and then find that I never use it for any of a number of reasons.

Heck, lots of folks swear by EMACS. Me, not so much. Similarly, lots of folks seem to like PC-Lint, but maybe they're all those danged EMACS users ...

--
Rich Webb     Norfolk, VA

That is not true you can try it for free on the web site

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

In message , 42Bastian Schick writes

Quite so. What price the Nikon D300 amateur camera? About 1500 USD? Lenses 500 USD upwards....

Hobby people often spend a LOT more than professionals on their hobby.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

In message , Rich Webb writes

It's not.

The alternatives cost an order of magnitude more so PC-lint IS the cut price version of static analysers.

One thing I do know is a major multinational that has a moratorium on spending paid many thousands of USD for its annual support for their static analysers.. They actually have the figures on how effective it is which is why in times of crisis it is one of the areas they will spend money.

In the professional end of the market where they have done the studies they find static* analysis is essential for good quality software. Apparently Open Source programmers can't see the point of it. I rest my case :-)

• my spell checker changed it to "satanic"! :-)

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

If so, that's new at least for the windows version. I haven't updated to the latest yet but the previous version wasn't node locked.

Robert