[pmwiki-users] Request: Transparent PNGs in PmWiki

Allister Jenks arj at zkarj.co.nz
Mon Mar 20 03:39:03 CST 2006

On 3/20/06, Joachim Durchholz <jo at durchholz.org> wrote:
> Note that my actual point is that it's simply too complicated and
> fragile in general. You may be able to make it work for me, and for
> everybody else who's coming around, but it's simply not worth the
> hassle: there will always be somebody with a strange IE version or a
> different plugin version or disabled JS who won't see what you intended
> to show. It's simply too many permutations.

Well, it doesn't use Javascript, but I take your point.  But then I am happy
that several 'mainstream' versions of IE (and those are the most prevalent)
display correctly.  I know it's a kludge, but I'd like to think the majority
will see what I want them to.

KISS is the principle to follow here, I think. These IE hacks are too
> complicated, take too much code to work, are too dependent on specific
> browser versions and parsing quirks (I'm pretty sure that most of the
> known "IE hacks" will not Do the Right Thing when IE 7 enters the scene.
> Not to speak of all those browsers that fake their browser
> identification strings.)

I normally agree with KISS whole-heartedly.  Yes, IE7 will be interesting.
I might need to extend the hack to  say  'but not if...'.

I'm quite happy with the script I posted earlier - it overlays the
> transparency on the server side and sends simple nontranparent PNGs,
> which is what will work even on the most braindamaged browsers. (I guess
> they built PNG transparency support into IE 7. In that case, the issue
> should go away in five years or so...)

It's a pretty smart idea, but unfortunately will not work in my case.
Because I am rendering over a tiled background image I have no (easy) way of
knowing what the offset of the PNG will be in relation to the background.  I
suppose it could be calculated but it would involve parsing most of the page
structure and even then probably would need browser specific code, as box
models are not all created equal!

I'll stick with what I have for now, but thanks for your input.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/pmwiki-users/attachments/20060320/f4645e45/attachment.html 

More information about the pmwiki-users mailing list