Bug 312280 – WebKitLegacy: printing a WebView embedded in an enclosing NSPrintOperation drops all text

• Status: RESOLVED
• Resolution: FIXED
• Component: Printing

[Jonathan Hammer]

304253@main (5f8a1e5) routed WebHTMLView's _setPrinting: through LocalFrame::setPrinting unconditionally. The embedded-print path (-[WebHTMLView adjustPageHeightNew:...], used when the WebHTMLView is a subview of a larger view tree being printed) calls _setPrinting: with all page dimensions equal to zero. Before 304253@main, WK1's layout helper explicitly gated forceLayoutForPagination on a nonzero minPageLogicalWidth and fell back to forceLayout(). That gate was lost in the refactor.

With zero pageSize, shouldUsePrintingLayout() still returns true for the top frame, so forceLayoutForPagination runs with a zero page, which asserts in resizePageRectsKeepingRatio in debug and produces a degenerate paginated layout in release — line boxes collapse and text runs are never emitted to the CG print context, while raster content (images, backgrounds) still paints. Users see print/export output with images but no text.

[Jonathan Hammer]

Pull request: https://github.com/WebKit/WebKit/pull/62742

[Radar WebKit Bug Importer]

<rdar://problem/174756900>

[Jonathan Hammer]

Created attachment 479070 [details]
Visual example of buggy behavior

Just to visualize what this bug is referring to. Attached is a PDF of a WebView exhibiting the buggy behavior (text is not rendered and layout is wrong).

[Jonathan Hammer]

Created attachment 479071 [details]
Visual example of fix

This is the same web page, this time with the patch applied. Text and layout is now correct.

[Jonathan Hammer]

Requesting that this pull request please be landed via the merge queue. The PR was code-reviewed and passes EWS for everything (except a spurious ios-wk2-wpt failure). This bug is impacting our shipping app and we’d love to see the fix land in WebKit. If the PR is deficient in any way, please let me know. Thank you!

[Simon Fraser (smfr)]

It looks like the latest push got no EWS coverage (perhaps it was down at the time). Could you rebase and update the PR please?

[Jonathan Hammer]

Yes, no problem. Thank you. I updated the PR last night and ran EWS. All green except for win-tests (two failures, both of which seem unrelated).

[EWS]

Committed 312629@main (d70314bc51fa): <https://commits.webkit.org/312629@main>

Reviewed commits have been landed. Closing PR #62742 and removing active labels.