Re: [Nautilus-list] Re: gtk theme compliance

[Darin Adler <darin eazel com>]

on 1/4/01 1:02 PM, Nicolas Mailhot at Nicolas Mailhot email enst fr wrote:

> I know. I've been doing it on the gtk font front. That's not
> fun:(.

Ramiro Estrugo is working hard on the font part right now. Many font
problems should be fixed on his next big check-in.

But I haven't noticed many bug reports from you about fonts -- just one
broad one that said many sizes are hard-coded, another specific one about
sizes hard-coded in the first-time druid, and a third that asks for a
more-flexible way to set the sizes used under icons.

But I routinely find hard-coded fonts and font sizes in the source code in
places that no one has reported. I would understand if you don't have the
time or the desire to help us with this, but if you do want to help, I
assure you that writing specific bug reports for specific places that are
broken will make a difference.

I'm sad to notice that unlike most people who report bugs, we've not fixed
any of the 11 bugs you reported -- for others our track record is much
better (for example we've fixed at least 20 that menthos reported); perhaps
this is why you don't want to report more bugs

> Thought I'd do some preventive action for once.

Complaining on nautilus-list, especially in a side remark in a thread about
something else, is probably not an effective preventative action.

> But you really should make mandatory for your coders to
> check their parts with one of the more garish gtk theme out
> there (using very big or very small fonts). That would take
> them half a sec (just the time to change the gtk theme in
> the control center) and this kind of bugs would leap at
> them.

I understand your proposal.

There are many useful verification steps people can take after writing code.
You've described one that's particularly good for finding bugs of this kind.

> You could even take the time to write a dedicated gtk test
> theme with very basic colors (blue, green, cyan...) to this
> end.

Nice idea. We would love it if someone contributed this to GNOME. It's
unlikely that the Nautilus team will take the time out from 1.0 development
to make this, but I bet our test team and perhaps some of our developers
would use it if it existed.

> Much more productive than to have users write obscure bug
> reports and then process them at helix.

(Eazel, not Helix.)

I understand your opinion on this, but I don't agree.

It's very effective to have people who are affected by bugs report the
actual symptoms they are suffering, because that helps make it clear how
truly important a prompt bug fix is. If it's just a theoretical thing, it
will usually get lower priority than things that users are actually
complaining about.

> This kind of visual bug is hellish to describe, but really
> easy do detect if  you take a little time to challenge your
> color/fonts assumptions with a radically different theme.

I understand your proposal for how to test thoroughly for this kind of
problem.

I'd like to repeat my request for bug reports. We'll then be likely to fix
the problems that bother you most even if we don't choose to institute the
thorough testing protocol you've described.

    -- Darin