search

sumatrapdfreader / sumatrapdf

SumatraPDF reader
http://www.sumatrapdfreader.org
GNU General Public License v3.0
13.73k stars 1.73k forks source link

Problems with AZW3 files #1314

Open R-Usher opened 4 years ago

R-Usher commented 4 years ago

I added two azw3 files to my OneDrive here: https://1drv.ms/f/s!AKm4ZaWdL2tfgi8 File G1.azw3 can be displayed only with UseFixedPageUI = true, file G2.nospdf.azw3 cannot be displayed in any mode. Both are generated by the same software.

GitHubRulesOK commented 4 years ago

@R-Usher I am only commenting on the file structure rather than source or provenance. SumatraPDF has obviously been written to support those file formats that either are DRM free or in the case of PDF have a password for the user to open Any other files structure that may have been modified to include/circumvent DRM cannot be expected to be handled correctly. SumatraPDF can open some standard .mobi structures in a file with an AZW or AZW3 extension and I have suggested to @kjk that that the AZW# extension could thus on occasion be an acceptable one.

However it is up to the user to ensure there are no problems with any internal modifications. I occasionally will use an e-book verifier to check such files pass testing before raising any support issue. One simple test is to see if it opens in native MuPDF (current version 1.16) without problems. If it is unsupported by MuPDF then raise a bug report there.

kjk commented 4 years ago

Technically, the failing .azw3 file (G2.nospdf.azw3) has 29 records (out of 70+) that are incorrectly compressed.

I've added a heuristic to to only fail if more than half records fails to load, which makes this file load. And it looks like enough data is decompressed that it actually shows the content.

So this particular file is fixed but I'll keep this bug open to further investigate.

It might be that my decompression code is not correct.

Which software do you use to create those files?

Did you use different settings to create G2.azw3 and G2.nospdf.azw3?

Do you use other viewers for .azw3 files and do they show Go.nopsdf.azw3 file? If yes, which readers are those?

kjk commented 4 years ago

BTW: now .awz3 files will open in ebook mode if useFixedPageUI is not set.

You can now switch between useFixedPageUI via Debug menu.

R-Usher commented 4 years ago

Both G1 and G2 files are properly opened by old Calibre 1.48 (which doesn't know how to properly open Y.nospdf.azw4). All files were some editors' giveaways.

R-Usher commented 4 years ago

Build 11628 displays both G1 and G2 in both ebookUI and FixedUI modes. In both cases SumatraPDF displays only cover picture, other embedded pictures (JPEG/JFIF) aren't displayed. And on the last page there is some embedded CSS code displayed. Old Calibre 1.48 displays all pictures and no CSS code.

GitHubRulesOK commented 4 years ago

It is very much down to how such files are converted here they are built in mobi format by 3.12 Getting Started with Data Structures.zip

Calibre does not do AzW3 output well but AzW3 to ePub is also reasonable Getting Started with Data Structures epub.zip