Mozilla full of CSS rendering bugs?

comments (AT) probertencyclopaedia (DOT) com (Matt Probert) writes:

Is it just me (probably) or is Mozilla and the newer Firefox full of
CSS rendering bugs?

I expect it is. But the ones you list below seem unlikely.

I ask, because some strange effects occur with Mozilla and Firefox
which don't happen in Opera and dare I say it, IE.

Like, for example, non-selection of a font-face for one style element,
even though it is selected okay for another element and the colour,
decoration and size are all rendered correctly!

Not recognising the 'border' parameter (eg: border: 1px solid black

Not rendering 'width' correctly (eg: width: 180px

This is where some test case URLs would be really useful. There could
be hundreds of reasons, only one a browser bug, why these don't do
what you expect.

--
Chris

Is it just me (probably) or is Mozilla and the newer Firefox full of
CSS rendering bugs?

I ask, because some strange effects occur with Mozilla and Firefox
which don't happen in Opera and dare I say it, IE.

Like, for example, non-selection of a font-face for one style element,
even though it is selected okay for another element and the colour,
decoration and size are all rendered correctly!

Not recognising the 'border' parameter (eg: border: 1px solid black

Not rendering 'width' correctly (eg: width: 180px

Matt

--
If your encyclopaedia doesn't list "widget glass", you're reading the wrong encyclopaedia.
The Probert Encyclopaedia. Its not the same.
http://www.probertencyclopaedia.com

"Matt Probert" <comments (AT) probertencyclopaedia (DOT) com> wrote

Is it just me (probably) or is Mozilla and the newer Firefox full of
CSS rendering bugs?

I ask, because some strange effects occur with Mozilla and Firefox
which don't happen in Opera and dare I say it, IE.

Like, for example, non-selection of a font-face for one style element,
even though it is selected okay for another element and the colour,
decoration and size are all rendered correctly!

Not recognising the 'border' parameter (eg: border: 1px solid black

Not rendering 'width' correctly (eg: width: 180px

Are you comparing apples with apples, e.g. by ensuring that a DOCTYPE is
used that triggers standards mode in all these browsers?

Examples would be nice.

On Thu, 15 Jul 2004 16:50:45 GMT "C A Upsdell"
<cupsdell0311XXX@-@- (AT) XXXrogers (DOT) com> broke off from drinking a cup of
tea at to write:

Are you comparing apples with apples, e.g. by ensuring that a DOCTYPE is
used that triggers standards mode in all these browsers?

No, which is why I'm asking! I had wondered if it was to do with
DOCTYPEs, which I shall now look into. Thanks for the suggestion - i
did ask if it was just me! <g>

Matt

--
If your encyclopaedia doesn't list "widget glass", you're reading the wrong encyclopaedia.
The Probert Encyclopaedia. Its not the same.
http://www.probertencyclopaedia.com

On Thu, 15 Jul 2004 18:46:33 GMT comments (AT) probertencyclopaedia (DOT) com
(Matt Probert) broke off from drinking a cup of tea at The Probert
Encyclopaedia to write:

On Thu, 15 Jul 2004 16:50:45 GMT "C A Upsdell"
cupsdell0311XXX@-@- (AT) XXXrogers (DOT) com> broke off from drinking a cup of
tea at to write:

Are you comparing apples with apples, e.g. by ensuring that a DOCTYPE is
used that triggers standards mode in all these browsers?

Pkay, it was a matter of DOCTYPES. Thanks for pointing me down that
road.

I thought CSS was fixed. There's CSS1 and CSS2 and they are defined,
but changing the DOCTYPE results in *very* different interpretations
of the CSS suggestions.

Perhaps I should wade through the incomprehensible brambles that are
the W3C specs <g>

Matt

--
If your encyclopaedia doesn't list "widget glass", you're reading the wrong encyclopaedia.
The Probert Encyclopaedia. Its not the same.
http://www.probertencyclopaedia.com

On Fri, 16 Jul 2004, Matt Probert wrote:

Pkay, it was a matter of DOCTYPES. Thanks for pointing me down that
road.

I thought CSS was fixed.

That's almost correct. (Just teeny adjustments in CSS2.1, and a few
things taken out of CSS2 that it seemed nobody was going to implement
- but nothing remotely on the scale of the pratical differences
between CSS2 specification and actual browser behaviour).

There's CSS1 and CSS2 and they are defined, but changing the DOCTYPE
results in *very* different interpretations of the CSS suggestions.

Now you're confusing buggy client behaviour with differences in the
specifications.

Perhaps I should wade through the incomprehensible brambles that are
the W3C specs <g

Not really. Browser "quirks modes" are an attempt by the browser
makers to perpetrate their original bugs (so as not to upset those
foolish virgins who designed their pages "to" the browser bugs,
instead of staying within the specifications) - while still offering a
standards-conform-wards mode for properly made pages.

Once they'd taken that decision, they needed some kind of handle for
guesstimating whether a page was intended to be standards-conforming
or reliant on old bugs. Unfortunately, they seem to have hit upon a
rather unfortunatate choice of handle for that purpose, namely the
notorious "DOCTYPE switching". The rest, as they say, is history.
But you won't really find any of this in W3C specifications.

*Brian* <usenet3 (AT) julietremblay (DOT) com.invalid>:

Alan J. Flavell wrote:

Unfortunately, they seem to have hit upon a rather unfortunatate choice
of handle for that purpose, namely the notorious "DOCTYPE switching".

Your posts hints that there might have been another (better) way.

To use an *HTML* feature to switch *CSS* rendering behaviour, is really
strange. If they would at least parse the HTML differently, more correctly
in "standards conforming" mode!

Was there?

If, I would have used the approach of using standard compliant mode by
default and as soon as an parse error occured, the rendition would restart
in quirks mode. This would of course make buggy pages take longer to load,
which is not such a bad thing IMO.

--
The penis mightier than the sword.

"Christoph Paeper" <christoph.paeper (AT) nurfuerspam (DOT) de> wrote

Unfortunately, they seem to have hit upon a rather unfortunatate choice
of handle for that purpose, namely the notorious "DOCTYPE switching".

Your posts hints that there might have been another (better) way.

To use an *HTML* feature to switch *CSS* rendering behaviour, is really
strange. If they would at least parse the HTML differently, more correctly
in "standards conforming" mode!

Was there?

If, I would have used the approach of using standard compliant mode by
default and as soon as an parse error occured, the rendition would restart
in quirks mode. This would of course make buggy pages take longer to load,
which is not such a bad thing IMO.

The problem with this is that you can have a valid page that depends on
quirks mode.

I personally would have preferred it if a META tag had been used instead a
DOCTYPE to specify the mode.

On Fri, 16 Jul 2004, Brian wrote:

Alan J. Flavell wrote:

standards-conforming or reliant on old bugs. Unfortunately, they
seem to have hit upon a rather unfortunatate choice of handle for
that purpose, namely the notorious "DOCTYPE switching".

Your posts hints that there might have been another (better) way. Was
there?

You put me in an awkward position there, since to give a good answer
to that, I'd have to adopt the position in which perpetuating legacy
bugs for the sake of wrongly-designed pages was a good idea. I can
sort-of understand the motivations that makes a browser vendor feel
that they have to do that, but frankly, I'm sick and tired of seeing
people saying "I know that my page is fine because it works in IE, but
it fails in [Mozilla/Opera/whatever]". And 99% of the time, the fact
it does what the misguided author intended in IE proves only that it's
been designed to rely on specific bugs and error-corrections provided
by IE.

But IE have been known to fix bugs, and to change their error fixups,
whereas the interworking specifications (for all that they may be
flawed in parts) do provide a firm framework for proper interworking.

I recall a number of cases from the early days of Netscape. As each
new version came out, some "feature" of the previous version upon
which web page authors had been relying would get fixed, and then
authors would scramble to fix their web page defects. A few weeks or
months later, it was all over, and people wondered what the fuss had
been about.

But now we have seemingly thousands of millions of broken pages out
there, and little prospect of most of them getting repaired, hence the
push to maintain buggy browser behaviour indefinitely. I can only
deplore this.

So, I have very mixed feelings about the implementation of quirks mode
at all. But assuming we want it in spite of that, how would we go
about designing it? An open-ended question. Obviously, we can't
expect these legacy wasters to put any kind of explicit flag on their
pages saying "please can you render this in quirks mode", even if
we're hoping to promote standards-conforming web pages into the future
by having our browser "default" to standards-conforming behaviour. On
the other hand it would be rude to demand that standards-conforming
authors should put some additional "real-W3C" declaration onto their
pages in order to get standards-conforming mode.

So I suppose it's inevitable that the switching has to be done by some
kind of indirect clues as to what the authors intended. OK, the
bottom line seems to be that I haven't a better solution, I don't want
there to be a solution (other than "fix those misbegotten pages"), but
I'm still grumpy at the solution which was picked. Does that answer
your question? (no).

all the best.

Brian wrote:

Matt Probert wrote:

it was a matter of DOCTYPES.

Perhaps I should wade through the incomprehensible brambles that
are the W3C specs <g


You'll find nothing on doctype switching afaik. Doctype switching was
an idea created by the browser makers, ostensibly to accomodate badly
authored pages but introducing a new set of problems. Some of the gory
details:

http://gutfeldt.ch/matthias/articles/doctypeswitch.html

Matthias suggests that IE and Moz introduced Doctype Switching when
AFAIK Mac was the first to use it, followed by NS/Moz, then IE and
finally Op.

--
Gus