cpdf in.pdf -gs gs -gs-malformed-force -o out.pdf [-gs-quiet] appears to scrub text from the resulting pdf

[greggarson]

trafficstars

Seeing this occur with cpdf 2.3 after I upgraded and tried to improve my malformed pdf handling.

[johnwhitington]

Are you able to supply the file in question?

Could cpdf previously handle it on its own?

[greggarson]

I can supply the offending file where would be the best place to send it to? cpdf would handle the file correctly previously but I don't believe the cpdf in.pdf -gs gs -gs-malformed-force -o out.pdf offering was available before 2.3? Or if it was we were unaware of it and not using it.

[johnwhitington]

You can send it to me at john at coherentgraphics.co.uk.

Normally, you would only want to use -gs-malformed, which tells cpdf to try mending a file with gs if it cannot mend it itself upon loading.

The only time you would want to use -gs-malformed force is when cpdf /can/ load the file ok, but the malformity causes cpdf to fail halfway through some operation. Thus, we would need to force the file to be repaired by gs beforehand.

[greggarson]

I see, let me test with that approach, I had assumed that -gs-malformed-force was an all powerful fixer, given that it took into account page level issues where -gs-malformed may not?

[johnwhitington]

No, the actual operation is exactly the same. In some future version, we might be able to enable cpdf to fix the file with gs even if the problem is discovered after loading the file, but that's not possible yet -- hence the existence of -gs-malformed-force.

[greggarson]

Update on this, I was running the command above as part of a script and not alone on the command line so wasn't looking at the output of that one command but rather the end result.

When actually running the command by itself I end up with:

bash-4.2$ cpdf test.pdf -gs /usr/bin/gs -gs-malformed-force -o out.pdf
Command line must be of exactly the form
cpdf <infile> -gs <path> -gs-malformed-force -o <outfile>

So it seems that the command doesn't even run when the invocation is as requested?

[johnwhitington]

It works here. Try just 'gs' instead of /usr/bin/gs?

[greggarson]

Seems to have the same result

bash-4.2$  cpdf test.pdf -gs gs -gs-malformed-force -o out.pdf                                                                                                                                           
Command line must be of exactly the form                                                                                                                                                                          
cpdf <infile> -gs <path> -gs-malformed-force -o <outfile>                                                                                                                                                         
bash-4.2$ which gs
/usr/bin/gs

[johnwhitington]

Fascinating. I'll send you a new executable with some debugging output so we can see what's actually happening.

You are using Linux/x86-64 right?

[greggarson]

That I am.

[johnwhitington]

Please try

http://www.coherentpdf.com/14thDecember2019.tar.bz2

This should print some debug output to help me understand why the special command line recognition for -gs-malformed-force is not working.

[greggarson]

Will do, thanks again for the help.