From: Daniel Stenberg (Daniel.Stenberg@sth.frontec.se)
Date: Sat Feb 06 1999 - 16:41:50 CST
On Sat, 6 Feb 1999, Ashley M. Kirchner wrote:
> > Does anyone have any objections against us starting from my 2b13 version to
> > move on?
>
> Yeah, I do, together with everyone else who were complaining about
> several of the beta ones.
So there shouldn't be any bugs in the released betas, or at least no that
crash hard, is that the message?
> They crash with a SEGFAULT, generating minimal core files, and or
> generating next to nothing for core that cannot be traced either way you
> try to figure out why it's crashing.
They do? They *all* do? Wow. How come no one told me?
Btw, I don't control the size of the core. And memory clogging bugs very
often corrupts the stack and then you get a hard traced bug. That's just
facts of life, or at least software developing. Things to fight and solve.
> This has been the case for a while, and for several people who have
> reported this problem.
Yes, some people reported about the problems back in November or so, since
then I've released three patches that recently lead to my 2b13 version. Not
one single person have reported anything on those corrections.
> And everytime, the same answer came back: 'Well, it doesn't crash on my
> machine, or I cannot reproduce it. Let's just go on.'
I never used those words. I never thought of the problems that way. I also
tried all reported problems on more than one machine, on more than one OS.
> Sorry, that's not enough. This is also why a lot of people stuck with
> the 2b3 version and are (eternally) waiting for Kent's b4 version. Why?
I've never claimed my versions were more stable than b3. My versions are and
were more like work-versions.
> Because his works without problems.
They do??? Where have you been all this time? You missed all mails from
people that get problems because they receive mails with long lines? "Kent's"
version have static buffers all over that are easily ran over.
All the way until b3 and even b4 if I read Kent's comments correct, used the
old and very silly assumptions all over that no mail line was ever longer
than 1024 bytes, and lots and lots of sub functions were using that same
assuption.
My goal with my work from that point was to remove all those limits. I knew
it would be a tough job, and I thought I was specific about that in my
release mails. I was detailed that for stability you'd better stay with b3.
> Yes, I also believe in work being done. What I don't believe in is
> in frequent, cripple releases. Or at least, if they're frequent, they
> work on fixing the problem before continuing on other matters.
So where were your help? Where were the developing team that should bring
hypermail 2 forward? You actually blame me for not being able to fix all
problems I knew people had?
When developing a program, you don't need "fixing the problem before
continuing on other matters". You can still fix item B even though you know
item A has problems. Even though fixing both is the prefered thing to do.
> > I'd love to be a part of an active team developing this further to a perfect
> > product!
>
> Current problems needs to be addressed before continuing with others.
Feel free to do so. I sure wouldn't mind more peers that eye through source
code and hunt for mistakes.
Unless there's a bunch of others with your hostile attitude towards me, I'm
sure gonna do my best too.
> And if something cannot be resolved because it can't be reproduced
> elsewhere, then we have a problem.
We have problems. They're called bugs. We have never seen anything but
beta-versions of hypermail.
> I never released any of my apps without knowing that it works 100% on all
> the supported platforms it was meant to work on. I never favor one over
> the other.
Hypermail isn't released. And if you're so good, how come you haven't helped
us out yet? We do need help as you have spotted.
I'd also like to point out that already version b3 contained a lot of code
from me.
Now go flame someone else please.
--
Daniel Stenberg - http://www.fts.frontec.se/~dast
ech`echo xiun|tr nu oc|sed 'sx\([sx]\)\([xoi]\)xo un\2\1 is xg'`ol