369362 - <use> broken for <symbol viewBox=...>

Status: RESOLVED FIXED

(Core :: SVG, defect)

Description

[jonathan chetwynd]

reversion

Comment 1

[jonathan chetwynd]

no wonder I couldn't find 'reversion' in the keywords
cheers

Comment 2

[Robert Longson [:longsonr]]

This works for me on Windows XP:

http://www.w3.org/TR/SVG/images/struct/Use01.svg

Comment 3

[jonathan chetwynd]

Attachment: <symbol viewbox=...>

Comment 4

[jonathan chetwynd]

please find a reduced testcase attached

Comment 5

[tor]

This is likely a dup of bug 369402.

Comment 6

[jonathan chetwynd]

#5

Tor, please recognise you have a problem and come to terms with it.

this bug was filed before bug 369402

Comment 7

[Jonathan Watt [:jwatt]]

peepo, you're not really doing yourself any favors by making remarks like that to the owner of the code you want changed. as far as the word "dup" goes, it doesn't matter so much which bug came first. if a bug filed later on has had more work done on it (and that could be diagnosis work, not just coding work) then it's quite normal to mark earlier bugs as dups. you shouldn't take that as a personal slight on you or your bug filing. I don't see any problem here, either with tor or with this bug. tor was simply making a note for future reference so we can remember to check whether any future fix for bug 369402 has fixed this one.

Comment 8

[jonathan chetwynd]

be as that may, tor has marked my reports as invalid on more than one occassion. When they clearly were not.

this bug has a reduced testcase attachment the other does not.
they appear to be dupes.

Comment 9

[jonathan chetwynd]

Attachment: viewbox width & height scaling broken

Comment 10

[jonathan chetwynd]

I don't fully understand viewbox, however I do use it for scaling.

the second attachment demonstrates the problem:
sites are sub-miniature and might be visible when zoomed ~:"

compare with opera or webkit for instance.

Comment 11

[Jonathan Watt [:jwatt]]

tor didn't close this bug. he only made a comment for future reference. regarding other bugs, it's not in our interest to close bugs that are valid. whenever we do so it is because - rightly or wrongly - we genuinely believe that they are.

I'm guessing part of the reason for your annoyance is that your bugs seem to frequently end up in close-open-close wars (we just had bug 317270 for instance). This is every bit as annoying for us as it is for you, since we'd rather get on with fixing clear bugs than repeatedly have drawn out battles to explain reasons and spec details. (Also reopening a bug that one of the devs has closed is essentially a rather aggressive way of saying "you're wrong!" to the people who spend a lot of time reading specs, and you've said yourself you're not a spec expert. Once or twice is fine, but when it becomes a regular it begins to grate.) At the same time I appreciate it's frustrating to have the bugs that are important to you closed for reasons you don't understand or agree with. As I've said before, you should feel free to continue discussion on closed bugs _without_ reopening them. We will sometimes have misunderstood or made a mistake, so we don't stop reading your comments just because the bug is closed and we'll certainly reopen it if you convince us that we're wrong. I think that small change on your part would smooth your interaction with Mozilla devs somewhat. ;-)

The fact that you take time to write small testcases is appreciated by the way. That's almost always a great timesaver when figuring out a bug, and we're now starting to put more emphasis on automated testing for which we need to have testcases for each bug.

Comment 12

[jonathan chetwynd]

#11

jwatt, your comment would have greater merit if tor hadn't marked https://bugzilla.mozilla.org/show_bug.cgi?id=369054 as invalid, pointing to the specs as evidence, but apparently not having read them.

Comment 13

[Jonathan Watt [:jwatt]]

What do you mean by "apparently not having read them"? As far as I know the spec does not say that you are right and tor is wrong. While I think our implementation is less useful than Opera/Safari, it seems to be a perfectly valid interpretation of the specification. It all hinges on the definition of "character cell", so it seems to me the validity of the bug is still in question. The bottom line is that when the spec is unclear tor gets to make the judgment call as owner of the SVG code. (Personally I'd rather we think about changing things in this case.)

Comment 14

[jonathan chetwynd]

#13 jwatt I'm taking this back to the bug as we are off topic for this bug....