1714217 - Investigate header/footer alignment inconsistencies on OS X

Status: NEW

(Core :: Printing: Output, enhancement, P4)

Description

[Emily McDonough [:alaskanemily]]

This results in the extra fuzz values on the test for bug 1713404

Comment 1

[Daniel Holbert [:dholbert] (PTO until Jun 29)]

If you've got a Try run with a log that shows this sort of failure (e.g. perhaps from your recent work in Bug 1730091), it might be nice to attach the log (or a snippet of it that includes the failure screenshots), to help with understanding this bug in the future.

(where the future understanding might be for the purposes of fixing this, or revisiting this to see if it's been fixed vs. if we're still seeing the same issue, or seeing if new similar-seeming failures are actually this bug or are something completely distinct).

Comment 2

[Emily McDonough [:alaskanemily]]

Attachment: test.png

Comment 3

[Emily McDonough [:alaskanemily]]

Attachment: reference.png

Comment 4

[Emily McDonough [:alaskanemily]]

Attachment: bug.log

Comment 5

[Emily McDonough [:alaskanemily]]

The relevant portion of the log:

[task 2022-02-03T01:04:47.058Z] 01:04:47     INFO - TEST-FAIL | layout/base/tests/chrome/test_printpreview.xhtml | different: 84, maxDifference: 172 
...
[task 2022-02-03T01:04:47.086Z] 01:04:47     INFO - Buffered messages finished
[task 2022-02-03T01:04:47.087Z] 01:04:47     INFO - TEST-UNEXPECTED-FAIL | layout/base/tests/chrome/test_printpreview.xhtml | Printing data:text/html,<!-- runTest47 --> and printpreview_bug1730091_ref.html should produce the same results - got false, expected true

The header/footer is slightly offset compared to the actual margins of the page.

Comment 6

[Daniel Holbert [:dholbert] (PTO until Jun 29)]

Thanks! Yeah, looks like the header and footer are positioned slightly further away from the top and bottom edges of the page, in the testcase, as compared to the reference case.

(It's several pixels of distance -- it's not e.g. an antialiasing artifact or something subtle like that. I wonder if this represents some actual measurable offset that we're applying, which we should be accounting for in our reference case here. Maybe some bonus unwritable margin, or something like that.)