1717805 - Missing data from table

Status: VERIFIED FIXED

(Firefox :: PDF Viewer, defect, P1)

Description

[Raluca Popovici, Desktop QA]

Attachment: table.pdf

Tested with:
Nightly 91.0a1 (2021-06-22)

Tested on:
Win 10

Preconditions:
In about:config, set pdfjs.enableXfa = true

Steps:

1. Launch firefox
2. Open the attached PDF

Actual result:
No data in table is displayed

Expected result:
A table with data should be displayed

Comment 1

[Raluca Popovici, Desktop QA]

Attachment: Nightly.png

Comment 2

[Raluca Popovici, Desktop QA]

Attachment: Adobe.png

Comment 3

[Calixte Denizet (:calixte)]

There are 2 issues in this pdf:

• the data (from dataset entry) are not correctly got because the root subform has no name when the root from data has one.
• the elements in table are rendered thanks to tb layout but a keep element prevents the table to be splited on several pages but because that stuff is no correctly handled we've a infinite loop (which is broken at some point to avoid it).

Anyway the first issue is addressed in:
https://github.com/mozilla/pdf.js/pull/13618

Comment 4

[Marco Castelluccio [:marco]]

Attachment: Link to GitHub pull-request: https://github.com/mozilla/pdf.js/pull/13618

Comment 5

[Marco Castelluccio [:marco]]

This is still problematic, the only content in the PDF now is the American Film Institute Lists text.

Comment 6

[Alin Ilea, Desktop QA]

Attachment: 2021-06-29_15h59_27.png

Hi,
Also here the most of the text is missing.

Comment 7

[Calixte Denizet (:calixte)]

:Alin Ilea, it's likely another problem.
Could file a new bug (same product/component) and attached the pdf ?

Comment 8

[Alin Ilea, Desktop QA]

@Calixte, sure, here is the new bug for comment 6 issue:
https://bugzilla.mozilla.org/show_bug.cgi?id=1718725
Thanks!

Comment 9

[Marco Castelluccio [:marco]]

Attachment: Link to GitHub pull-request: https://github.com/mozilla/pdf.js/pull/13598

Comment 10

[Daniel Bodea [:danibodea]]

I can confirm the fix in Nightly v91.0a1 from 2021-07-07.

The table.pdf document displays even better than Adobe Reader.
(FYI comment 3, the table is properly split into pages, also.)