Acrobat DC Pro stopped working with BPDX files (again)

Report · Oct 24, 2022

The BPDX feature of Acrobat Pro has stopped working (again?). This is the same issue reported a year ago in https://community.adobe.com/t5/acrobat-discussions/adobe-bpdx-not-working-after-latest-updates/m-p/1....

I'm running Windows 10 Pro 21H2 OS Build 19044.2130 and Acrobat Pro Continuous Release Version 2022.003.20263.

I have a Powershell script started by Windows Scheduler every weekday morning at 1am. The last action is to dete the PDF index log file and index folder (leaving the .pdx file), then run Acrobat with a BPDX file to rebuild the search index. This has worked perfectly for a year or more. The last successful run was Tuesday 18 October. Since then the log file shows the process hangs at the step of running Acrobat with the BPDX file.

I've checked and confirmed that all required files and folders are present and correct.

If I manually build the index by selecting Index from the Tools pane, click Full Text Index with Catalog, open the existing index definition file and click Build, the index is built with no errors.

As an experiment, if I open a Command Prompt and enter:

"C:\Program Files (x86)\Adobe\Acrobat DC\Acrobat\Acrobat.exe" C:\Flare Tools\PDFIndex.bpdx

the Acrobat starts and displays an Information message "There was an error opening this document. This file cannot be found." followed by a Critical message "C:\Flare Tools\Tools\PDFIndex.bpdx could not be opened for reading". So Acrobat has munged the path to the BPDX file (it is C:\Flare Tools\PDFIndex.bpdx), presumably because it contains spaces.

If I then run Acrobat with the proper quoting:

"C:\Program Files (x86)\Adobe\Acrobat DC\Acrobat\Acrobat.exe" "C:\Flare Tools\PDFIndex.bpdx"

Acrobat starts, there is no error message, but nothing happens. The index folder and log file do not appear in the folder where they should be.

If I then try to manually build the index (as described above), Catalog runs but displays the error "Could not open document: <document-name>" for every file that it should be indexing.

If I then double-click on one of the PDF files in Windows Explorer, Acrobat opens the PDF file without error.

If I then try to manually build the index (as described above), everything works as it should.

Report · Oct 25, 2022

Hi,

Since you've shared a link that describes the same issue after an update, I am curious if the file associations changed after an unattended update took place on that machine (wether it was an update of the Acrobat Pro program or the operating system itself) .

Just to rule out other things, Would you be able to test the same script in another computer with the same specs and see if it behaves in the same way?

But going back to my theory of a possible change with file association, I've seen similar behaviors after updating in more recent versions of Windows 10 and mostly on Windows 11; specifically after Internet Explorer 11 reached end of support back in June.

With that in mind, Microsoft Edge browser could be acting as the file explorer and also defaulting as the PDF handler after each update... just food for thought.

If that would be the case, what happens if you completely disable IE11 with all its add-ons accelerators and extensions that used to integrate with Adobe Acrobat's file explorer and enable Internet Explorer mode on Microsoft Edge?

And also, what happens if you compare the same test when you execute it from the Powershell shell versus the old school command line prompt shell?

I would verify if any batch invoking commands have deprecated after an unattended update, and even check if the output file changed from text to a any type of UTF output file (just guessing in the dark here).

Report · Nov 07, 2022

I have investigated the file associations and browser behaviour you mentioned. The file associations are correct. The appropriate registry entries can be seen in RegEdit, and when I right-click on a bpdx file in Windows Explorer, the Open With command includes the option Adobe Acrobat DC. The Microsoft Edge/IE11 aspects are not relevant to the setup in our company, according to our IT department. The issue is still occuring with version 2022.003.20263. The issue also occurs on my home PC. I believe this is a regression issue like the issue reported a year ago.

Report · Nov 07, 2022

Interesting!

I'm digging up some info.

Report · Nov 07, 2022

So if you review this guidance:

https://helpx.adobe.com/acrobat/using/creating-pdf-indexes.html

Look if relatvie folder paths changed, if the type of font that was used to name the folders have changed to a font type that contains characters that the Index catalogue doesn't map, or also check if the batch script that is currently in use is trying to rebuild an entire index when it is not necessary instead of purging and preserving the original PDX file... just a guess in the blind.

Report · Nov 08, 2022

Thanks Rob for your continuing input. As I said in my original post, I've been using a PowerShell script run from Windows Schedular for over a year. One of the actions of the script is to regenerate the PDF search index by running Acrobat with a BPDX file. Its been running perfectly for over a year, then suddenly stopped after 18 October. This is one compenent of an established publication process and I have not made any changes to the system. Also, as I noted, I can regenerate the PDF index manually using the Index tool in Acrobat. It's just the BPDX aspect that is now failing.

Let me address your specific comments:

I am familiar with the link you provided. I used the information more than a year ago to create the setup I have. All the settings are correct and have not changed.
There has been no change to paths, relative or otherwise, as this is part of a larger publication system.
All folders were created using either Windows Explorer, Command Prompt or Powershell. I don't know how you could name a folder with anything else but the system default font. And again, nothing has changed since the process was working just fine.
The powershell script deletes the index log file and the index folder containing the last built index, leaving just the PDX file. It then runs Acrobat with the BPDX file to build a new index. Again, this has been working perfectly up to 18 October.

Earlier, I found another post reporting similar problems for an Acrobat release that post-dated the post from a year ago (https://community.adobe.com/t5/acrobat-discussions/adobe-bpdx-not-working-after-latest-updates/m-p/1.... In trying to locate this later post for my response, I found this link (no idea why it didn't turn up in my earlier searches) dated 1 October: https://helpx.adobe.com/nz/acrobat/kb/known-issues-acrobat-dc-reader.html. It lists the following:

"Problem: [Protected Mode: On] Unable to load the BPDX file in Acrobat with Protected mode turned ON. [4325240]"

This seems to resolve the problem. If I uncheck Enable Protected Mode, then BPDX files are processed. If I check Enable Protected Mode, then BPDX files are NOT processed. Adding privieleged locations does not seem to have any positive impact. If I check Enable Protected Mode, and add trusted locations for all my PDF, PDX and BPDX files, then BPDX processing still doesn't work.

Report · Nov 08, 2022

+++ EDITED REPLY, fixed many typos

Well, that is good to know.

With this discovery we can confirm that there is nothing wrong with your batch script, nor a bug that was introduced with the Index tool per se.

Now we can narrow down the issue to the enhanced security that is enforced via the Protected Mode in Acrobat.

Just for troubleshooting purposes, It would be a good idea to single out one of the PC's at work (or test with your PC at home), and use the Event Viewer to review the Application and System logs that are dated back for 18-19 October and compare them against the logs that are dated 16-17 October.

By comparing these logs you may be able to notice discrepancies with other software that interacts with Acrobat (which could've impact what you had working before).

For example, Windows unattended updates, Microsoft Edge web browsers updates, Microsoft Office365 upgrades (if applicable), operating system optional updates, C++ and .Net redistributables version updates, security updates, antivirus definition updates. MIME extensions, etc.

And if you do notice any discrepancies that could indicate a potential conflict, would you be able to bring back your test computer back to a restore point before the issue that popped on or after 18 October?

I already gathered a lot of info for you, but before I post back with my findings, I need from you to confirm if you were able to spot anything relevant in those logs.

Specifically, I will have to reiterate once again to pay close attention to anything that, indeed, could be related to Microsoft Edge and Internet Explorer 11.

Both web browsers enforce its own Protected View and sandboxing default security that in the past used to interact with Acrobat.

There are well-documented bulletins in Microsoft support and Adobe Help Center that indicate potential conflicts due to that integration.

As a matter of fact, you can test if Acrobat will behave the same when executing the BPDX with Protected Mode enabled in Acrobat while switching off "Protected View" for the Internet Zones in both Internet Explorer 11 and Microsoft Edge web browser.