FileMaker 14 v2 Crash-a-palooza?

[James Gill]

Just installed the FM14 update to version 2 today and I have gone from maybe 1 or 2 crashes per day to FileMaker crashing every 10 minutes now.  Anybody else encounter a drastic increase in the number of times FileMaker crashes daily?

[Wim Decorte]

Had no crashes before and none after upgrading to v2.   Does it happen with all files, even newly created ones?

[James Gill]

No, happens on an existing file that, prior to v2, would crash maybe once or twice a day

[Josh Ormond]

I haven't had any crashes with 14.

Is it all files, or specific ones, or specific layouts? The only possible common thread I have seen with anyone reporting crashes had to do with button bars and pop-overs. But, I use both, and haven't had any crashes.

[Rick Whitelaw]

James, a filé crashing once or twice a day still a huge nimber of crashes. Something must be amiss with the file. If a file of mine crashes it's a very rare évent. I WorldSource tory to isolante thé source of the crash within the file. Does it happen when running a script? During simple data entry? While switching layouts?

my two cents . . .

[Mike Duncan]

What platform and version of FM? Any plugins installed? Is the crash report indicative of display related problems? Are you using themes or is this a converted file from an earlier version?

[Ocean West]

I have seen this dialog several times. I am not sure that is a FileMaker dialog or a El Capitan dialog. 

Today I opened a hosted solution and had both the layout manger opened and the script work space open and as soon as I toggled close a folder in the scripts FMP crashed. It happened 3 times. The last time I closed the layouts manger, and scrolled the scripts list and closed folders with fewer scripts, finally managed to close the folder with lots of scripts. 

 

[Rick Whitelaw]

Sorry for my previous post. Apparently I had French selected as the language on my iPad so it mangled the crap out of what I typed.

[James Gill]

I narrowed it down to an ExecuteSQL query statement.  The really odd part is that the crash only happened if I called the statement just after calling a 'Delete portal row' script step.  I narrowed it down to a relationship in the query that was using a mis-keyed relationship (number to text).  Fixing the relationship stopped the crash from happening.

It's such a specific set of steps to reproduce the crash that I thought I might have found a bug, but I was unable to reproduce in a clean file.  Oh well.

[Josh Ormond]

I would still report that one James. That is definitely a bug. Glad you found it.

[James Gill]

In my experience, unless you can reproduce the issue in a clean file (i.e. either brand new or one of their demo files) FileMaker is not interested in hearing about the issue.  Remember the now confirmed PDF corruption bug?  I reported that to FileMaker close to 3 months ago but they said that if I couldn't reproduce it outside of my solution then it was a file/layout corruption issue.

[Ken Newell]

Both myself and my co-worker have been getting a lot of crashes using FMPA.  Simply scrolling through a script or debugging a script seems to be what is causing the crashes.  I have not tried creating a new clean file to see if I can reproduce the crash.  Testing  FMP with our users who do not use the script work space or debugger, without any reported crashes as of yet.  Need to get to the root of this before we can push this out for production.

[James Gill]

I've got a few select users running on FMPA 14 and, other than the PDF corruption bug, they haven't reported any adverse effects.  I move on to a wider FMP14 test tomorrow, using a few 14-specific features.  I'm holding my breath to see how it goes.

[JerrySalem]

A minion of mine installed the 14.02 update on his Mac.  Now he is telling me that the Keychain won't save his credentials anymore.

Every time he has to put in his UserName and PW even though he keeps checking the 'Save to Keychain' checkbox.

Anyone else get this result?  I am hesitant to try it out myself.

 

Jerry

[James Gill]

I had a similar issue for a little while but to be fair to FileMaker, I think it was KeyChain just being stupid.  KeyChain having issues?  Never... /s

Edited August 26, 2015 by James Gill

[James Gill]

Here's a doozy.  I have a Production server running FMSA 13 and a development server running FMSA14.  I run the same script sequence on both servers with the following results:

Test 1) FMPA13 with FMSA13 = No crash
Test 2) FMPA13 with FMSA14 = No crash
Test 3) FMPA14 with FMSA13 = No crash
Test 4) FMPA14 with FMSA14 = Crash

I narrowed it down to a specific script that was originally set up to GTRR (say about 20-25 records) and then perform a 'Delete All Records' script step.  This was causing a freeze & subsequent crash 100% of the time.  I changed it to a Loop with a Delete Record/Request script step and it no longer crashes 100% of the time, just when I don't use script debugger to loop through the records manually.  There are potentially many child records below the parent record that have cascading deletes enabled.

Here's the deal.  At this point I am completely fed up with FileMakers lack of support.  I know that if I call in then I am going to be told that my file is corrupt and that there's nothing they can do to help diagnose the crash.  "How do I know that," you ask? Because literally every time I have called FileMaker to report some bug or crash, that's what I get told.

Does anybody have ANY contact at FileMaker that can help with FM14 and the crashes that I'm experiencing?

Edited August 26, 2015 by James Gill

[JerryRobin]

A minion of mine installed the 14.02 update on his Mac.  Now he is telling me that the Keychain won't save his credentials anymore.

Every time he has to put in his UserName and PW even though he keeps checking the 'Save to Keychain' checkbox.

Anyone else get this result?  I am hesitant to try it out myself.

Jerry,

You probably already know this, but there is a new File Options checkbox in FM14: Allow Keychain Access to save password. Did this somehow get unchecked in the v2 update? (Of course, if it was unchecked, the option to Save to Keychain shouldn't even appear in the password dialog. Hmmm.)

- Jerry

Of course, you can probably ignore my comments, since according to FMForums I am a newbie.

[James Gill]

Well this is interesting. Turns out it's not a crash but a severe performance hit when using FMP14 vs FMP13, I just wasn't waiting long enough for the entire process to complete.
I tested FMP13 vs FMP14 with a file running on FMSA14 the same sequence twice for each client.

FMP13: 21.84 seconds and then <12 seconds to complete the script sequence (The time difference is I assume due to some caching on FMP's end)
FMP14: 3 minutes 27 seconds and then 3 minutes and 28 seconds to complete the same sequence

More and more curious...and here's a server graph of remote calls/sec comparing the two.

[JerrySalem]

Jerry,

You probably already know this, but there is a new File Options checkbox in FM14: Allow Keychain Access to save password. Did this somehow get unchecked in the v2 update? (Of course, if it was unchecked, the option to Save to Keychain shouldn't even appear in the password dialog. Hmmm.)

- Jerry

 

Of course, you can probably ignore my comments, since according to FMForums I am a newbie.

Jerry

 

Turns out a reboot fixed the issue.

 

Good suggestion though.

 

Don't worry about being a newbie with a little experience you might go far

Jerry

[Ken Newell]

Just a follow up to my earlier post about crashes . . . I was able to track it down to an outdated plugin.  The plugin crashed the debugger and script work space.  Updated the plugin and everything is working just fine. 

[dysong3]

I was going to add my grain of salt but just realized that the crashing was due to a script invoking an obsolete application.  Pay no mind.

Edited November 3, 2015 by dysong3

[James Gill]

So far V3 seems to be much more stable than previous versions.  Now that I have upgraded to v3, fixed font issues and upgraded to El Capitan, I maybe experience a crash or two per week.  A definite improvement.

[dysong3]

Now here's my grain of salt.  Since upgrading to El Capitan and to Filemaker Pro 14.0.3 a script containing "perform Applescript" continually makes FMP crash.  The script is

tell application "X-Lite" to activate
tell application "System Events"
try
keystroke "v" using {command down}
keystroke return
on error
tell application "X-Lite" to activate
keystroke "v" using {command down}
keystroke return
end try
end tell

What's weird is that in debugging mode it does not crash.  I use this script to dial a VOiP phone (X-Lite).  Would anyone have any ideas for a work around?  Something that can tell FMP to copy a phone number and send to X-Lite to dial.

[dysong3]

Answering my own question here.  I found a great script for phone dialing with X-Lite here https://www.youtube.com/watch?v=6YQJCEZQu94&feature=iv&src_vid=ALKK33DKa94&annotation_id=annotation_70984727

[oakbridge]

I found this thread after going through an afternoon of crashes and odd behaviour (Specify Field dialog box opening up as an empty window with just the header as an example). I had wondered whether others were seeing similar behaviour. 

Then I remembered that I had been running in 32 bit mode (Mac OS 10.10.5). I had been working last week on another project which used the FM-MoneyWorks connector. 

Switched back to 64 bit mode and (knock wood), FMPA has been very stable for the rest of the afternoon. 

So to emphasize what others have been saying, try looking at different parts of your environment when you see crashes. James I would not expect to see FM crash even the one or two times a week that you are experiencing. 

Edited November 8, 2015 by oakbridge

[James Gill]

On 8/26/2015, 5:01:23, James Gill said:

Well this is interesting. Turns out it's not a crash but a severe performance hit when using FMP14 vs FMP13, I just wasn't waiting long enough for the entire process to complete.
I tested FMP13 vs FMP14 with a file running on FMSA14 the same sequence twice for each client.

FMP13: 21.84 seconds and then <12 seconds to complete the script sequence (The time difference is I assume due to some caching on FMP's end)
FMP14: 3 minutes 27 seconds and then 3 minutes and 28 seconds to complete the same sequence

More and more curious...and here's a server graph of remote calls/sec comparing the two.

Coming back to this issue...this turned out to be a layout related issue.  The layout I was calling the script from had a rather large, complex portal on it.  I discovered that deleting the records away from that portal drastically improved performance, and adjusted the script to accommodate.

[Josh Ormond]

Glad to hear you at least found the cause of the problem in 14. Can you tell if it was just data/relationship or if it was something from the layout object itself?

[dysong3]

I came across another one of my scripts that includes an Applescript and which makes FMP 14.0.3 crash.  And once again it didn't in debugging mode.  I tried adding a 0.5 second pause just before the Applescript and that stopped the crashing.  Hope this helps.