[Nautilus-list] Re: Darin doesn't work effectively with Yoann?

[Yoann Vandoorselaere <yoann mandrakesoft com>]

Darin Adler <darin bentspoon com> writes:

> On Sunday, June 17, 2001, at 03:28  AM, Yoann Vandoorselaere wrote:
> 
> > Let's take the last one, that was also sent to the list,
> > about the re-layout being done even before any file is drawn.
> >
> > This is causing latency on directory load.
> >
> > I pointed you to the cause of the problem :
> > The hook to the files_changed signal in fm-directory-view.c that
> > isn't disconnected at the good time.
> >
> > I think this is specific enough... no ?
> 
> Not really specific enough. You've pointed at a problem area, but not
> any correct way of improving performance. I need to consider deeply to
> figure out how to change things so they still work but fix this
> problem area.

What I pointed you in this area, *is* a bug. A bug affecting performance.
The fix might not be OK, but the bug is still there.
 
> So I don't know what kind of response you're expecting. Yes, this is a
> problem area. What is your specific question again? Is your patch OK?
> No, it's not. I answered that in detail.

Yes you did. Now I'm gonna fix it for real.

> >> and you complain when I don't answer
> >> fast enough or you don't like the tone of my responses. Other
> >> contributors work with me in ways that are easier on me, and which
> >> take less of my time and effort to deal with.
> >>
> >> But I'll try to answer the ones that I can figure out when I have
> >> the time.
> >
> > Darin, this answer is easy, and I think that some of your arguments are
> > out of context. I asked you several time if something was wrong with me
> > because I wasn't getting any answer from you, to questions that date from
> > a long long time.
> 
> If the answer is "easy" then why am I the only one on the list who can
> answer it? This is not a one-man project, and I'd like you to figure
> out how to work with other contributors rather than complaining that I
> don't work with you fast enough.
> 
> > Working on Nautilus performance problem is not something easy, and the
> > only thing that you do when not at least acknowleding my question is
> > like saying "I don't care him contributing".
> 
> No, you ask for a lot more than that. You ask for quick answers to
> your questions, 

I don't think you can call > 2 weeks (should be 3 now) quick.

> and you also ask for respect and credit even when your
> changes are incorrect 

Your hypocrisy is really pissing me off (we had a similar 
discussion in private, and you completly change your mind
in public).

Tracing the performance problem take several hours, 
it's especially hard when you don't have the big picture.

Then you found one, you try to fix it. You send Darin a patch.

And Darin answer you : 
your patch isn't good, but I fixed the problem myself.

(And several time, not a damn credit in the ChangeLog for at least
 finding it / analyzing it).

Cool... it took several hours to trace this stuff, and when you are
on the way to fix the problem, you just feel like you're screwd.

cool. really.

Darin, I think you come from the commercial world and that you don't
have a really large experience with free software. Now let me tell 
you something : credit is probably the most important thing.


> and I have to redo them (and you get angry that
> I don't work with you so that you can correct them), 

Darin, you piss me off [again].

You should learn about respect and communication. And you should probably
learn about what Maintaining a free software project mean.

When someone send you a patch, that what the patch want to do is OK, but
the patch itself is not OK, you don't fix the patch yourself, but send a
mail to the author telling him what he need to change.

If you don't do so, 
no one will learn a damn lesson about the way to do thing.

Here is why I got angry :
I got angry because I spent *hours* tracing Nautilus performance issue.
I don't have quantify, and gprof isn't suitable for profiling Nautilus...
Do you know what this end with : printf assertion.

So it take *hours* to trace potential performance problem that way, 
and even thought you start to see one, it'll take much more time to get 
the whole picture.

> and you also ask
> for the "proper attitude".

Ok. Let's go for rude communication : next time you piss me off, 
I'll say you fuck off instead of trying a good way to solve the 
communication problem : nice.

I didn't asked for proper attitude, I asked you to be less offending. 

> > I've spent several night trying to audit Nautilus on the performance/
> > latency problem. This is not easy, Nautilus isn't simple.
> >
> > And this is kind of offending when you work for several hour on something,
> > that you send a mail to the maintainer asking patch review, or how
> > to avoid
> > X thing to happen or whatever question about a problem and never get any
> > answer...
> 
> You questions are relentless and extremely demanding. If I answered
> them all I would have no time to do anything else. So I really don't
> mind offending you a bit. If you had fewer questions and followed
> fewer false trails, I would probably be more enthusiastic about
> working with you more.


[quoted words are subjective]

The learning curve for Nautilus is very high, 
I myself start to have a "good" view of "how it work", but it took time. 

Now, in the beginning, it's true I asked many question, because
getting the "big picture" alone in such complex program would be very 
hard...

I also want you to know that I'm not a Gnome programmer nor a GTK 
programmer. Last time I did a GTK program was several years ago. I'm 
more working with low level stuff.

So you understand that Nautilus wasn't easy to learn for me.


>   But you tend to send out patches that effective turn off big pieces
> of the code that you don't understand and say "Is this OK?"
  
Yes, when you don't have the big picture, you sometime think something
will not cause problem in other place. But it will.

Now each time I fixed something, there was a reason to fix it, and
audit behind. Even if the fix wasn't ok, and raising several issue.

-- 
Yoann Vandoorselaere | Murphy's law : If anything can go wrong, it will.
MandrakeSoft         | O'Tool's commentary : Murphy was an optimist.