1953903 - Fix tests/pdf_viewer/test_add_image_pdf.py

Status: RESOLVED FIXED

(Mozilla QA :: STArFox, defect, P1)

Description

[Tracy Walker [:tracy]]

tests/pdf_viewer/test_add_image_pdf.py is failing on DevEd due to proper prefs not set to enable alt text, which is enabled by default in Beta/Release. Let's for it to work by setting the follow:
browser.ml.enable: Set to true (to enable local inference engine)
pdfjs.enableUpdatedAddImage: Set to true (to enable updated add image with alt text flow)

Comment 1

[Tracy Walker [:tracy]]

fixed in https://github.com/mozilla/fx-desktop-qa-automation/pull/521

Comment 2

[Tracy Walker [:tracy]]

Failing again in Fx137Beta9

Comment 3

[Tracy Walker [:tracy]]

(re)fixed in https://github.com/mozilla/fx-desktop-qa-automation/pull/534

Comment 4

[Tracy Walker [:tracy]]

This is failing again in CI/Beta and local runs

Comment 5

[Tracy Walker [:tracy]]

I couldn't figure out why this isn't working. However, I discovered that it's method, using pynput, for entering the location of the image file is leaving the picker dialog with "data" as default directory.. This test should clean up after itself and return default to "Downloads." That said, other tests like the csv export tests shouldn't assume the picker dialog defaults to anything in particular. Instead, those tests should set the location by entering full path, like this add image test does.

Comment 6

[Tracy Walker [:tracy]]

Ben made these changes in https://github.com/mozilla/fx-desktop-qa-automation/pull/555

Thank you!