Forum moved here!

Home / Pattern is not rendered properly

IvarSnaaijer

I tested Case 1571 of PDFium in Sumatra 3.1.2 64-bit and the same problem happens.
Only one box appears, while Acrobat shows both.

GitHubRulesOK

Agreed appears to be a MuPDf regression since 2017 !
There is a slim posibility the revision yesterday may have undone that issue but will have to wait for next SumatraPDF merge to confirm it works same as it did in 2016 ?

However (even then) neither did show the first unbound pattern only the boxed one !!
So thats a seperate bug that may warrent a visit to bugzilla.

image
imageimage

SumatraPeter

Perhaps an issue with aliasing and/or line widths? With Sumatra 3.2, at 644% zoom I get the same result, but at 645% both boxes render properly for me as far as I can tell.

GitHubRulesOK

Interesting I did not connect it with the semi related “asterix” issue where zoom & alpha was the problem

For me the break point is 577% to see both !!

SumatraPeter

Most likely due to the differences in our displays and graphics driver/OS settings - resolution, DPI, scaling and what not. Anyway at least the issue has been narrowed down and it’s confirmed that MuPDF isn’t failing to render the pattern completely. Hopefully the MuPDF devs can fix this as quickly as they tackled the Asterix one I encountered.

GitHubRulesOK

Hmm I wonder since SumatraPDF pre-rel is not working at any zoom level if there has not been a secondary regression since 3.2 ?

@kjk
May be worth waiting and looking at this one after the next MuPDF Lib update since yesterdays changes to alpha handling ?

GitHubRulesOK

@IvarSnaaijer
@SumatraPeter
looks like the latest SumatraPDF update with MuPDF (which seems to included the fix for Bug 702970 - mupdf renders PDF incorrectly ) did NOT resolve this issue

SumatraPeter

If this ^ is the bug you’re referring to, it seems to describe a different issue; so why would a fix for it resolve this one as well? I’d recommend that @IvarSnaaijer should report these rendering bugs he’s finding directly to Artifex so that upstream MuPDF fixes can be incorporated into Sumatra eventually.

IvarSnaaijer

reported @MuPDF, number 703044

SumatraPeter

also reported at Sumatra under case 3429

@IvarSnaaijer: Good job, although you might want to leave an additional comment there with actual links to this thread and the Chromium bug. I’m saying this because “case 3429” probably won’t help them find this discussion, as this is the Sumatra forum and the correct place to report bugs specific to Sumatra is via GitHub.