Re: [ace-users] gperf

[Douglas C. Schmidt]

Hi Dave,

> I am at a turning point again in my refactoring / redesign oportunities.
>
> I have identified that we are using some sort of 'gperf' process in the
> building of our product. We build all of the apriori known names in
> the product before shipping it into a perfect hash lookup scheme.
> It turns out that our pcode uses the actual hash values as aliases
> for the name, saving us additional symbol tables. I.e., we require
> 'minimal perfect'.
>
> This part of the code is ancient, dating back to 1990. I want to replace
> it with a modern version of perfect hashing.
>
> The question is: "Is the output of gperf considered GPL-ed?".

No, it's not.

> I noticed that you use and ship a version of 'gperf' with ACE. We do
> use ACE; however, we do not use TAO. Does the use of 'ace-gperf'
> carry a different license for its output?

No, but it shouldn't matter because the output of gperf is not GPL'd!

> Are there other commercial packages that use 'gperf' where they
> do not consider their program GPL-ed and do not deliver source
> code?

TAO's IDL compiler uses gperf and it's not GPL'd (it is open-source, but
with a BSD-style license).

> Are there other perfect hash generators with perhaps less restictive
> licensing for their output that I should consider?

Not that I know of, but it shouldn't matter since the output of gperf
isn't GPL'd!

Take care,

Doug