Re: Serious performance problems (malloc related?)

[Yitzchak Scott-Thoennes <sthoenna at efn dot org>]

On Sat, May 28, 2005 at 02:40:31AM -0400, Christopher Faylor wrote:
> On Fri, May 27, 2005 at 09:29:54PM -0700, Edward Peschko wrote:
...
> For the record: I do not give any serious consideration to the two or
> three email voices here who think they speak from some sort of moral
> high ground and are compelled to tell me how to communicate.  I do very
> much appreciate that you haven't lapsed into profanity or sarcasm,
> however.
> 
> 99% of the time when people report a problem they are saying "I have a
> problem.  I didn't go to too much effort to figure it out.  I didn't do
> any research about how I should report it.  I want you to help me now."
> My response to this kind of email is normally to point people to where
> they need to go to either fix the problem themselves or provide the
> information that we need to fix the problem.  If people don't like the
> way I do this, here's something that should be obvious by now: I don't
> care.  You can take my response or leave it.
> 
> In this case, the OP provided a test case, which is rare (and
> appreciated).  They didn't provide any other details other than that,
> although he did offer noise information about how he'd rather be using
> linux and how this was a serious problem.  He also offered a completely
> uninformed guess.
> 
> We normally ask for the guidelines from http://cygwin.com/problems.html,
> to be followed, however, I know that issues of performance are rarely
> cut and dry (although knowing the version of cygwin that was exhibiting
> problems would have been interesting) and so the real key here is to
> learn enough about the DLL to diagnose the problem.
> 
> What you interpret as "snippiness" is just cut and dry advice: If you
> have a lot of people relying on a product and the product is misbehaving
> then it makes sense to either purchase support or learn enough about it
> so that you can support it.  Relying on volunteer email to fix something
> which is impacting your whole organization is not a reliable way to get
> "serious problems" fixed.
> 
> FWIW, the original message violated many of the tenants of Eric Raymond's
> "How to Ask Smart Questions" and I responded just like Eric Raymond
> predicted.
> 
> >There are two ways of seeing it - everybody (or a huge subset of everybody) 
> >is wrong and I'm squeaky clean, or yes I could improve my attitude in how
> >to deal with people. 
> 
> One thing you may notice in my email, is that I normally try to I try to
> tell people what they need to do to get results.  I don't normally tell
> people that "everyone" thinks a certain way and I don't often make
> personal observations about people's character.  I just offer (often
> terse) advice on what I think someone should do to solve a problem.
> 
> As it turns out, for the most part, the way that I conduct myself
> provides me with the results that I want.  The mailing list is fairly
> self policing these days.  People who report "cygwin not work" bugs are
> directed towards the right way to report a problem and, often, their
> problems are fixed.  People who want to contribute are pointed at
> cygwin-apps or cygwin-patches and often they end up contributing.
> 
> Other people who want to change things but don't have a clear idea about
> what they want to do and only know that they are angry because I'm not
> nice to them, and won't give their half-formed thoughts the
> consideration they know they deserve, eventually go away.  Which is
> exactly what I want to happen.
> 
> This isn't a democracy.  It's just a successful free software project.
> I (and Corinna) get to make the rules and I (we) get to comport myself
> how I (we) see fit.  If you don't like that then you can take cygwin
> source code and make your own fork.
> 
> (And, boy will I be fuming.  That should offer some incentive at least)
> 
> >I know this cultural issue has deterred me in the past; I keep hoping that 
> >things'll change and I (and possibly many others) would change their mind
> >and start assisting cygwin.
> 
> I have to confess that I'm *glad* that you are deterred.  IIRC, in your
> communication to the cygwin list you rarely demonstrated a real grasp of
> the technical issues that you were trying to grapple with.  So, I'm
> happy not to have to either witness or correct your misperceptions on a
> regular basis.
> 
> Oh, and, there you go again.  First it was "everyone" and then it was
> "possibly many others".  I believe that this argument style is called
> "Appeal to Belief".  You've posted relatively rarely to the cygwin list
> and have never, as I mentioned, and as I recall, demonstrated any
> particularly strong grasp of either the technical or community aspects
> of cygwin so I don't accept your premise that you speak for "everybody"
> or "possibly many others".
> 
> Not that it would matter if I did.

IMO there should be a link to this message from
http://cygwin.com/acronyms/#WJM.

Comments?