Mojave - getting "pdftotext not found" error trying to compare PDFs

[luckman212]

BC4 - Version 4.2.6 (build 23150)
MacOS 10.14

Getting an External program "/Applications/Beyond Compare.app/Contents/MacOS/pdftotext" not found error when trying to compare 2 PDFs. Both files are in my "~/Downloads" folder. I have granted BC4 "Full Disk Access" as well in Security prefs. I also double checked that the "pdftotext" binary exists inside the BC app bundle (it does) and that it runs manually from Terminal (it does).



any idea? This was working as of last week...

[luckman212]

I figured out that I was trying to compare 2 "protected" PDFs. The error message shown in BC was misleading. When I ran the commandline tool on the files directly, I saw the real problem:

$ pdftotext foo.pdf
Permission Error: Copying of text from this document is not allowed.

I opened them in Preview.app, exported as unprotected PDF and voila, working again. I suggest though, maybe the error message can be made clearer so users know what's going on...

[Aaron]

Thanks. We've got an open issue about Mac OS PDF failure messages not reporting correctly in different scenarios. I'll add this case to our tracker entry on the subject.

[arvan]

I get the same error reported for a pdf file with parentheses in the name, can you add this case to the issue tracker if it's not already there.

[Aaron]

Thanks. Noted. As a workaround, if you add a space in front of the first (, or remove the () from the name, you should be able to get the conversion working.

[Chris]

Comparison of PDFs with () in the filename on macOS is fixed in Beyond Compare 4.4.5, just released earlier today.

[luckman212]

I just hit this again with BC 4.4.6 build 27483 and it had been so long since I posted this that I totally forgot about it...

May I suggest updating the bundled pdftotext library? The one that ships with BC currently is very old and outdated (2017). There is a fork that runs native on M-series CPUs and is quite a lot faster, called Poppler. It's available both via Homebrew (brew install poppler) or as a standalone: https://github.com/tsdgeos/poppler_m...lows/cmake.yml

I tried to add a custom format handler and point it to the Homebrew installed version at /opt/homebrew/bin/pdftotext but it complains "External program not found". I made sure BC has Full Disk Access. Not sure what's going on there.

[Aaron]

Hi,

Could we get a bit more info on the file that threw a conversion error? This thread is specific to file names with () that would cause a failure, which should be fixed as of 4.4.5 and 4.4.6.

You can also create your own custom File Format using any conversion utility. We have a guide here, using RESX as an example, but any conversion utility can be used:
https://www.scootersoftware.com/supp...rnalconversion

[luckman212]

Thanks for that tip. I was a little unclear how to set up the custom external File Format. That link was very helpful, I was able to link it to /opt/homebrew/bin/pdftotext and that's working nicely. I wish I took better notes on the specific files that were failing, but if it pops up again I'll come back and post them. I just did a quick test of some files with parentheses () and they worked with both the built in viewer as well as the external, so not sure what else was going on there.

I still suggest BC may want to integrate the poppler version above since it does run a lot faster on Apple Silicon.

[Aaron]

Hello,

It's something we can research, but requires digging into not just the application and code but the license if it can be included, provided as an extra download, or a user has to implement it themselves.

Sometimes the BC4 PDF error message itself can be a red herring; the most common issues when a PDF fails to convert are caused by Adobe security settings that prevent export, or the file is a picture of text and does not contain any actual text data. Both are quick to test using Adobe itself to open the file then attempt a Save As Text, and note what is generated (blank file?) or a warning or error.

[luckman212]

Still do recommend this if/when BC5 gets near...

[Aaron]

Thanks for the follow up. Swapping the PDF library isn't something we've been able to tackle yet, but we are getting closer with BC5. Hopefully we'll have more to share soon.