649114 - Text-anchor middle is not computed correctly for element scaled to 0

Status: RESOLVED FIXED

(Core :: SVG, defect)

Description

[Marek Raida]

User-Agent:       Mozilla/5.0 (Windows NT 6.1; WOW64; rv:2.2a1pre) Gecko/20110411 Firefox/4.2a1pre
Build Identifier: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:2.2a1pre) Gecko/20110411 Firefox/4.2a1pre

If text element, centered, is scaled to 0 and then scaled in, it behaves like text-anchor start and not middle. Having it initially scaled to very small, but non zero size - like 0.0001, helps, and it behaves properly then.

Reproducible: Always

Steps to Reproduce:
1. Open file from attachment

Actual Results:  
Text is after 1s zoomed in, right aligned to line in the middle (behaves like start anchor)

Expected Results:  
Text should be after 1s zoomed in, centered against line in the middle

All other browsers compute text position properly, even when initially scaled to zero

Comment 1

[Marek Raida]

Attachment: Broken testcase with initial scale = 0

Comment 2

[Marek Raida]

Attachment: Case with workaround, initial scale = 0.0001

Comment 3

[(mostly gone) XtC4UaLL [:xtc4uall]]

Confirmed against Mozilla/5.0 (Windows NT 5.1; rv:2.2a1pre) Gecko/20110412 Firefox/4.2a1pre ID:20110412030535

Comment 4

[Robert Longson [:longsonr]]

Attachment: patch

Comment 5

[Jonathan Watt [:jwatt]]

What is the SUPPRESS_INVALIDATION for? And why in the one case of setting updateGlyphMetrics but not the other?

Comment 6

[Robert Longson [:longsonr]]

Attachment: address review comments

Comment 7

[Jonathan Watt [:jwatt]]

Comment on attachment 590778 [details] [diff] [review]
address review comments

Thanks.

Comment 8

[Robert Longson [:longsonr]]

pushed https://hg.mozilla.org/integration/mozilla-inbound/rev/f500e9a0253c

Comment 9

[Robert Longson [:longsonr]]

That didn't go so well. Pushed https://hg.mozilla.org/integration/mozilla-inbound/rev/1efdc864bce2 to backout patch

Comment 10

[Robert Longson [:longsonr]]

Need to look into 

411334-1.xml
621918-1.svg
621918-2.svg

Comment 11

[Robert Longson [:longsonr]]

relanded https://hg.mozilla.org/integration/mozilla-inbound/rev/5313091609ad without the SUPPRESS_INVALIDATION part that seemed to upset things so much.

Comment 12

[Joe Drew (not getting mail)]

https://hg.mozilla.org/mozilla-central/rev/5313091609ad