Some failing tests produce visually identical output on my platform

[lornajane]

I added some good hashes for tests that "fail" but where the PDFs are exactly the same as the reference ones.

[ralsina]

Just as a sidebar: there are some cases where the test is not purely visual (for example, PDF TOC issues) perhaps those should be marked as requiring a manual check beyond the visual one.

[akrabat]

Just as a sidebar: there are some cases where the test is not purely visual (for example, PDF TOC issues) perhaps those should be marked as requiring a manual check beyond the visual one.

Good point. Do you mean cases such as links?

(It's so annoying that different versions of ReportLab generate different PDFs that are the "same"!)

[ralsina]

Yes, links, metadata, some other things I don't remember.

…

On Sun, Sep 30, 2018, 5:03 PM Rob Allen ***@***.***> wrote: Just as a sidebar: there are some cases where the test is not purely visual (for example, PDF TOC issues) perhaps those should be marked as requiring a manual check beyond the visual one. Good point. Do you mean cases such as links? (It's *so* annoying that different versions of ReportLab generate different PDFs that are the "same"!) — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#686 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAAGK0iqSPxKe6Eea5FWYqw9EaKqaVwyks5ugSN7gaJpZM4W1xUt> .