Using environment variables without leaking memory?

[Rich Felker]

On Thu, Oct 26, 2006 at 11:00:44PM -0400, Rob Landley wrote:
> No, the BSDs have been doomed from day 1 because their license makes it 
> commercially appealing to create a proprietary fork of the project if it gets 
> any momentum, and hire the developers away to work on said proprietary fork.  

Someone's equally free to fork a popular BSD and put the fork under
GPL. This is actually what should have been done a long time ago. BSD
license is utterly stupid, as you explained quite well.

> > Aside from the few 
> > abominations invented in committee, like pthread, aio, and some of the
> > more obscure regex requirements,
> 
> c++,

C++ is not in POSIX at all. Are you saying this is a bad thing?

> the gaping holes in the standard glossing over stuff like "init" 
> and "mount"...

There's no reason things like init or mount should be in the standard
since it does not benefit anyone for them to be common between
systems (except people making malware that sticks its nose in places
it doesn't belong and tries to alter the system's boot process behind
the administrator's back). The only things that should be standardized
are things that ordinary user programs will depend on.

> > C and POSIX are very good standards 
> > in that they standardize what was historical practice on traditional
> > systems when possible, and make minimal changes for the sake of
> > compatibility (which are usually easy to adopt into existing
> > implementations) when historical implementations disagree or when
> > there's an obvious deficiency.
> 
> I'm not writing code for you.  I'm writing code for me.  I put it up for 
> approximately the same reason my blog is public.  Telling me I shouldn't do 
> this means that I have an incentive to make sure you can't easily use my 
> work, assuming I care about you at all.

I'm not telling you you shouldn't do anything. And yes I have the
source and I'll be quite happy to use it. However I hope that the apps
that have nothing GNU/Linux-specific to them, like the shell, will be
portable and run anywhere, not just GNU/Linux. If not then I'll apply
whatever fixes are needed. However I'd rather not play the OpenBSD
game where the "upstream" for all the packages is highly OBSD-specific
and separate people have to maintain a "portable" version, like with
OpenSSH and (great program I recently discovered) mg.

> > Fortunately there are only a tiny number of programs that ever need to
> > modify the environment (sh and env are all that come to mind).
> 
> Hi, are you unaware I'm writing a shell from scratch?

Yes, and quite happy about it. Even if what you write is horribly
nonportable, you can't possibly make it more nonportable than Bash. :)
And I have confidence that you write very good and efficient code, so
it will be easy to adapt if necessary. Of course I hope I can just use
your upstream directly without having to make my own version of it.

> Let's ignore the fact that you've obviously never done CGI programming and 
> focus on this useless "command shell" thing that obviously does obscure stuff 

I never said any of it was useless, just that what it's doing is not a
common need of lots of applications and thus that the deficiency in
the standard, while annoying for anyone writing a shell, is rather
small in the big scheme of things.

Rich