Re: Someone using COMPAT_SVR4(_32) ?

[Dave McGuire <mcguire%neurotica.com@localhost>]

On 08/01/2017 12:42 PM, Maxime Villard wrote:
>> This is a really silly argument.  Solaris is still under active
>> development.
> 
> Yes, but COMPAT_SVR4 is *not* under active development. No one maintains
> it,
> and it is proven buggy, contrary to compat_sunos - I guess because no
> one has
> given a close look at it yet. If it is seen that compat_sunos is unused and
> unmaintained, then let's remove it too. But that's a different issue.
> 
> What exactly is your argument? That "someone" "may" use "some" solaris
> programs?
> But is there at least one binary that is solaris-specific, that is not
> available
> on linux and not open source? And if such binary exists, is it proven
> that it
> works with the current compat_svr4 implementation?
> 
> At some point, old, unmaintained code needs to be removed. What I'm
> trying to
> figure out is whether there's any real use case that justifies keeping this
> feature, even if disabled. So far, you haven't demonstrated anything.
> And if
> you're so passionate about this, why didn't you maintain compat_svr4 in the
> first place? Probably because you're not using it either.

  Please forgive me for jumping in.

  History has shown that, in the general sense, the parts of an
open-source project that are fun and flashy get most of the attention
from developers.  Writing new whiz-bang features to show off in the
release notes is a lot more fun than chasing bugs and maintaining
existing code, so that's most of what happens.

  When you couple that with the "this is buggy, just rip it out"
mindset, what we'll be left with is just the parts that developers find
it fun to work on.

  Ripping out good features because they simply need to be fixed is an
awful idea.

                 -Dave

-- 
Dave McGuire, AK4HZ
New Kensington, PA