Closed GoogleCodeExporter closed 8 years ago
I can fix the bug, but there'll still be the no-gradient issue until the link
is broken...would it make sense to convert symbol->svg on import for Firefox?
Original comment by adeve...@gmail.com
on 1 Oct 2010 at 7:39
Also that particular file didn't appear to work anywhere due to a missing
xmlns:xlink="http://www.w3.org/1999/xlink" in the root SVG...
Original comment by adeve...@gmail.com
on 1 Oct 2010 at 7:40
I think it would make send to convert symbol->svg and maybe not just for FF
since someone may import it into svgedit in a non-FF, save to their HDD and
then open the SVG directly FF (not through svgedit). I don't know - the other
alternative is just for FF and then perhaps the default behavior when saving as
SVG could be to convert svg->symbol. It wouldn't take care of direct viewing in
FF, but would allow for much better (not perfect) results in Illustrator and
still be ok if they opened again in svgedit in FF.
Does that make sense, or did I just go around in circles ;)
Original comment by adrianbj...@gmail.com
on 1 Oct 2010 at 7:53
I just downloaded from the library itself and it works fine in Illustrator and
has the xmlns:xlink
http://ian.umces.edu/imagelibrary/displayimage-4694.html
Does that work for you?
Original comment by adrianbj...@gmail.com
on 1 Oct 2010 at 7:54
Sorry, you are right, there is no xlink declaration, but it does still open in
Illustrator ok
Original comment by adrianbj...@gmail.com
on 1 Oct 2010 at 7:58
You know...there is a solution that I'd dismissed before for its potential for
leading to other problems, but thinking about it more it really might not be so
bad.
Symbols with gradients CAN work in Firefox...just as long as all the gradients
are directly under the <defs> and not inside the symbol. The concern there of
course is that it's less clear (when looking at the source) which gradient
belongs to which symbol, and so it's less obvious whether the gradient should
be deleted with a symbol, if it's deleted.
But...we could certainly try this (maybe just for FF), and see whether or not
that would actually lead to problems. So I think I'll first give that a shot
(next week, probably).
Original comment by adeve...@gmail.com
on 1 Oct 2010 at 8:18
I have added the xlink dec and it certainly loads fine now - my regex during
upload to the library will now take care of this for all future images - now I
just need to update all the existing ones.
Also, Chrome is not showing the image at all - just an empty and unscaled
bounding box.
I have changed the summary to reflect the real issue - hope that is not too
confusing.
Original comment by adrianbj...@gmail.com
on 1 Oct 2010 at 8:33
Interesting thoughts on placing the gradients directly under the <defs>. As
well as the issues for svgedit (which I am sure you can code to overcome), I
wonder what impact it will have in illustrator/inkscape when someone goes to
break link to symbol or delete a symbol - just another possible problem to
think about :)
Would be nice if it was workable though!
Original comment by adrianbj...@gmail.com
on 1 Oct 2010 at 8:37
So trying that in r1774, works okay in FF for this example. Note that breaking
the link of the final symbol causes the new group to shift in this case,
because the symbol has a viewBox set. Shouldn't be terribly hard to fix, just
something I wasn't expecting to see. :)
Looking into the Chrome bug now.
Original comment by adeve...@gmail.com
on 4 Oct 2010 at 5:16
Apparently the viewBox is what's causing Chrome to freak out...ugh.
Original comment by adeve...@gmail.com
on 4 Oct 2010 at 5:55
Thanks for all your efforts on this (and all the other cool things you have
been working on) and sorry for the day of silence - I am run off my feet at the
moment. Testing the trunk I can't get the Ruppia fruit to load at all in FF or
Chrome via image library import or open. Sorry for the lack of further info. I
think I am going to go through and "fix" all our images with a symbol wrapper.
Of course I think that svg-edit still needs to be able to handle these
regardless.
Original comment by adrianbj...@gmail.com
on 5 Oct 2010 at 9:12
Strange...both import and open work fine in Firefox for me.
Original comment by adeve...@gmail.com
on 6 Oct 2010 at 4:34
That is weird indeed - I finally got around to rebooting my computer and it is
certainly working fine now. I had tested several times before with browser
restarts, cache clearing etc and no dice. Anyway, looks good - thanks :)
Original comment by adrianbj...@gmail.com
on 7 Oct 2010 at 12:44
Oops - I think there is a nasty side effect - images without <symbol> are
"importing" all black in FF, but, they are "opening" ok. I just tested the
usual crab and duck examples. BUT, fine in Chrome!
Original comment by adrianbj...@gmail.com
on 7 Oct 2010 at 12:50
Fixed in r1780. I also tried opening one of these gradient-outside-symbol files
into Illustrator to see what would happen...the file opens just fine, and when
you save it again it stuffs the gradient element back into its symbol. :)
Original comment by adeve...@gmail.com
on 7 Oct 2010 at 5:43
Just noticed that when saving with FF 4 without the server save extension, the
symbols don't appear in the browser for me. Save Page As does result in a valid
svg though, so it seems that the <symbol> element is still having issues in FF.
Maybe resort to the webkit approach of copy/paste save?
Original comment by adrianbj...@gmail.com
on 10 Oct 2010 at 11:04
Sounds to me like the usual FF problem where anything defined in the defs (like
gradients) don't show up in data URLs. Should be a warning message about that
the first time you save.
Original comment by adeve...@gmail.com
on 11 Oct 2010 at 2:42
I am thinking that this can probably be closed, although perhaps the saving
issue in FF still needs to be addressed. I just wanted to make a note here
though that the Ruppia maritima example file mentioned in the initial report no
longer has a <symbol> element, in fact I have cleaned up all the symbol
elements from all the images in the IAN library. Probably wasn't necessary, but
just wanted to simplify things as much as possible.
Original comment by adrianbj...@gmail.com
on 12 Oct 2010 at 9:33
Closing as suggested.
Original comment by adeve...@gmail.com
on 12 Jan 2011 at 3:25
Original issue reported on code.google.com by
adrianbj...@gmail.com
on 1 Oct 2010 at 6:01