Applying Boolean build expression causes loss of data tagged with condition tags that are excluded as "NOT"

[ReneSteph]

Hi All,

I'm running FM12.0.2.389 on Win7-SP1 (64-bit) laptop with i7 & 16GB RAM.

I reported this issue as a bug in the Adobe bug base, but they weren't able to reproduce the bug with the files I sent. This leads me to believe that perhaps the conflict is a result of something gone wonky or conflicting in my system. Here's what's happening:

Using "Set as per Expression" triggers FM crash upon next File>Save, displays error about imported graphic, and results in loss of data that was tagged with conditions set to exclude via "NOT" in the Boolean build expression.

Steps to Reproduce Using My Files (zipped):

1. Open Ch1_Introduction-4FC.fm.

2. Create a build expression named 3DpartnerKitConditions using this definition:

not "LIVEonly" and not "InstallOnly" and not ( "Comment" or "Deleted" or "Future" or "international" or "PostInstall" or "Question" or "SDinstallOnly" )

3. File > Save.

4. Apply the build expression.

5. File > Save...(crash)

Result: All data tagged with conditions excluded via Boolean NOT is not longer in the file, as seen in Ch1_Introduction-droppedText.fm.

What I THOUGHT would happen is that the data tagged with conditions excluded via Boolean NOT should remain in the file but hidden, and could be shown readily via Show/Hide Conditional Text > Show All with Show Condition Indicators. Sometimes it FM doesn't crash and it does behave this way, but 90% of the time the result is the crash and loss of data.

Other things that I'm considering, I think I have some scripts or plugins running that were written for previous versions of FM. Yes, I'm addicted to stuff from CudSpan, Bruce Foster, Silicon Prairie, FrameExpert, Sundorne, and Miaramo. I have noticed that SafetyMIF (admittedly written to FM11) causes FM crash on exit, but since we're using Git repositories for incremental check-in, trying to produce MIFs routinely is quite a hassle otherwise, and checking in the FM bin file bloats the repository way too much for engineering to be at all happy.

I don't know what's triggering the Boolean tag expression issue, but I really, really need to solve it. I'm a solo writer (for now) with a very heavy load of docs and short deadlines that can only be managed by sound single sourcing methodologies. So, these files rely on variables, condition tags, shared files among different books, shared graphics, and text insets, as appropriate.

Can someone in the community help me come up with a way to pinpoint the root issue? I'm considering approaching the troubleshooting by first uninstalling FM, cleaning up the typical Windows uninstall trash, reinstalling FM fresh, then adding back the plugins one at a time and attempt running the Boolean tag setting after each plugin is added to see if I can tell which one is triggering the issue. If it's a FM installation gotten corrupted, that would solve the problem, too.  What do you guys think?

Thanks,

Rene

[Arnis Gubins · Accepted solution]

Rene,

It's a sound approach and probably about the only thing viable, since there may also be interactions amongst them (though I probably have most of the same ones and don't see any odd-ball crashes when saving).

The Bruce Foster ones are not compatible with FM12 (or even 11) - too old. Some of Chris's plug-ins also had issues with FM11 & 12 unless you had the correct 32-bit MS runtimes installed (even on a 64-bit operating system).

[ReneSteph]

Arnis,

Thanks as always for your insight. After reinstalling, I'll not reinstall Bruce's plug-ins.

How do I verify (and obtain if necessary) the correct 32-bit MS runtimes? If I can run SafetyMIF with a 32-bit runtime, I'll just "jones" for the rest. 

Rene

[Arnis Gubins]

The link Chris sent me was: http://www.microsoft.com/en-us/download/details.aspx?id=30679

Select VSU_4\vcredist_x86.exe at the prompt after clicking the download link.