1819335 - Extra blank sheets are sometimes inserted when `size: <orientation>` causes an orientation switch

Status: VERIFIED FIXED

(Core :: Printing: Output, defect)

Comment 1

[Jonathan Watt [:jwatt]]

Found this while debugging some Google Docs issues.

One example that demonstrates this is the testcase bug 1816554 (attachment 9317487).

Comment 2

[Daniel Holbert [:dholbert]]

Attachment: testcase 1 (just two named pages, with `break-after:page` also causing a forced break)

I'm not sure this actually has anything to do with size or orientation. I think it's just a case where we've got two reasons for a pagebreak -- a named page difference, and page-break-after:always (or break-after:page).

(Google Docs probably runs into it for page-boundaries-that-involve-a-size-change just because there's a difference in named pages at that point.)

Here's a reduced testcase. I think we might already have a bug filed for this, too; I know it's something we've run into in the past, with named pages and forced pagebreaks.

Comment 3

[Daniel Holbert [:dholbert]]

Attachment: frame dump of testcase 1

Here's the frame dump from testcase 1. It shows that page 2 just has a PageBreak frame and nothing else. We probably push it there due to the page-name mismatch logic.

Really, we should try to get the PageBreak frame at the end of page 1 (or omit it altogether, if we can tell it'll be unnecessary).

Comment 4

[Daniel Holbert [:dholbert]]

Attachment: testcase 2 (just two named pages, with `break-before: page` also causing a forced break)

For symmetry, here's a testcase like testcase 1 but now with break-before on the second element, instead of break-after on the first element. We should be sure that our fix handles both of these cases.

Comment 5

[Daniel Holbert [:dholbert]]

Marking this as blocking the gsuite-printing-metabug.

jwatt has a gdocs testcase https://docs.google.com/document/d/1N9kDNCdB-6YLW-Y0fLt_--9LyaN-hvFX4t-B6KoEOe0/edit where this seems to trigger a mostly-blank second page (which, incidentally, gets some weird double-printed overflow from page 1 -- I'll spin off a separate bug for that unexpected overflow).

Comment 6

[Daniel Holbert [:dholbert]]

(In reply to Daniel Holbert [:dholbert] from comment #5)

Marking this as blocking the gsuite-printing-metabug.

jwatt has a gdocs testcase [...] which, incidentally, gets some weird double-printed overflow from page 1 -- I'll spin off a separate bug for that unexpected overflow

--> Filed bug 1859634 on that.

Comment 7

[Emily McDonough [:alaskanemily]]

Attachment: Bug 1819335 Part 1 - Do not create a forced break for CSS named page change immediately before an nsPageBreakFrame.

Comment 8

[Emily McDonough [:alaskanemily]]

Attachment: Bug 1819335 Part 2 - Add tests for page name change and break-after/break-before at the same point.

Because the break-after/break-before should cause the new page name to be on
the next page, there should only be one page-break.

One of these cases is still not solved, but the simple case works now.

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/7f7ef196a61d
Part 1 - Do not create a forced break for CSS named page change immediately before an nsPageBreakFrame. r=dholbert
https://hg.mozilla.org/integration/autoland/rev/d171871eb9d2
Part 2 - Add tests for page name change and break-after/break-before at the same point. r=dholbert

Comment 11

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/42851 for changes under testing/web-platform/tests

Comment 12

[Iulian Moraru]

https://hg.mozilla.org/mozilla-central/rev/7f7ef196a61d
https://hg.mozilla.org/mozilla-central/rev/d171871eb9d2

Comment 13

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Upstream PR merged by moz-wptsync-bot

Comment 14

[Web Platform Test Sync Bot [:wpt-sync] (Matrix: #interop:mozilla.org)]

Upstream PR was closed without merging

Comment 15

[Ardelean Oana, Desktop Test Engineering [:oardelean]]

Reproducible on a 2023-10-27 Nightly build on Windows 10, using the testcases from Comment 2 and Comment 4.
Verified as fixed on Firefox 121.0b4 and Nightly 122.0a1 on Windows 10, Ubuntu 22, macOS 12.