Some progress with httpd on Busybox 1.1.3

[Rob Landley]

On Thursday 11 February 2010 06:49:11 Denys Vlasenko wrote:
> On Thu, Feb 11, 2010 at 7:21 AM, Rob Landley <rob at landley.net> wrote:
> > I do know a lot of breakage has since been fixed and cleanups made.
> >  (Back then httpd was maintained by an insane russian who posted and read
> > messages via translation software instead of actually learning english,
>
> Vladimir Oleynik.
> Actually I don't think he used a translator,
> he was genuinely try to speak English.

No, he used a translator program and made no attempt to learn english, ever.  
I confirmed this with him on several occasions, and that was the reason that 
after years of interacting with him on this list our ability to communicate 
with him never improved in the slightest.

  http://lists.busybox.net/pipermail/busybox/2001-January/035695.html
  http://lists.busybox.net/pipermail/busybox/2006-February/052845.html

The translation results sucked at his end too, which is why his "too long, 
didn't read" threshold was shorter than this email, so you _couldn't_ hold a 
serious conversation with him.

> He is just one of those people for whom learning foreign
> languages is _hard_. Not everyone is equally good at it.

I know.  I personally absolutely suck at it, and sympathize with people 
working with development communities in a language they don't natively speak.  
I try hard to make allowances for language issues, and I greatly respect the 
effort people are making.

But Vladimir wasn't actually making any effort.  Add in the posessiveness, 
mediocre code quality, unwillingness to discuss design issues, unwillingness 
to defer to the project's maintainer (not even when it was Erik Andersen)...

Also, keep in mind that he never actually wrote new software.  He just took 
other people's code and "ported" it to busybox.  Things like httpd, ash, and 
fdisk were all external projects he nailed on to busybox without much cleanup.  
But if anybody else tried to clean them up, he maintained his version out-of-
tree and did all further development against that, and the only patches he was 
willing to submit were a complete sync with his tree reverting anything else 
anyone had done with it...

It got old.

> > and freaked if anyone else touched "his" code.
>
> Unfortunately, it is not at all untypical behavior for many.

Yup.  I'm currently wrestling with the fallout from a previous maintainer of 
qemu's powerpc target:

  http://lists.gnu.org/archive/html/qemu-devel/2007-04/msg00198.html

Which is the reason that statically linked powerpc binaries still don't run 
under qemu-ppc three years later, because she coded to the register layout IBM 
specced (for AIX) instead of what the Linux kernel on powerpc was actually 
doing, and when this was pointed out insisted that the Linux kernel change its 
established behavior to match the spec (for AIX).

Rob
-- 
Latency is more important than throughput. It's that simple. - Linus Torvalds