If only we had known that earlier...



  • We've been trying to track down a bug in our product that is causing [i]another[/i] application on our customer's system to crash.  This is feasible, as our app interacts with this other app in ways that could possibly cause a crash.  We've been running around, chasing down every screenshot and crash report we could find, trying to reproduce the issue on our end, setting up conference calls, virtualized test machines, custom builds/hotfixes of our application...

    So, finally, after pleading for more information, we get a [i]tiny[/i] little detail from the customer.  Something so minor that we can understand why they wouldn't bring it up...

    The crash occurs when [i]our[/i] application isn't even running.



  • The mere presence of your app makes the other app crash.  Sounds like there are some serious issues to be resolved.



  • @purge said:

    The mere presence of your app makes the other app crash.  Sounds like there are some serious issues to be resolved.

    Agreed - dll-hell type issue? Registry/common-config issue?

    I had an issue about 15 years ago where if our (C-motif) app was running, all the dialogs on another app turned to black-on-black making them (somewhat) more unusable than usual. Turns out that both apps were modifying the X-lib global colormap, and foregrounding either app would obliterate the GUI components of the other (until you switched back).

     



  • Or perhaps it has absolutely nothing to do with their app, and the other developers are simply trying to pass blame? 



  • @purge said:

    The mere presence of your app makes the other app crash.  Sounds like there are some serious issues to be resolved.

    Actually no.  We've found the issue for them, and it involves application A causing application B to crash when using feature X.

    We are the developers of an unrelated application C. 



  • @djork said:

    @purge said:

    The mere presence of your app makes the other app crash.  Sounds like there are some serious issues to be resolved.

    Actually no.  We've found the issue for them, and it involves application A causing application B to crash when using feature X.

    We are the developers of an unrelated application C. 

     

    Sounds like you could be a colleague of mine. We get that all the time...



  • @djork said:

    Actually no.  We've found the issue for them, and it involves application A causing application B to crash when using feature X.

    We are the developers of an unrelated application C.

    I was actually personifying the applications and thinking perhaps that application B's moods are simply a result of unspoken anger or jealousy over application C.  Throwing feature X into the situation only complicates the matter.   



  • @djork said:

    @purge said:

    The mere presence of your app makes the other app crash.  Sounds like there are some serious issues to be resolved.

    Actually no.  We've found the issue for them, and it involves application A causing application B to crash when using feature X.

    We are the developers of an unrelated application C. 

    Bonus points if you got them to take you out for drinks after they sent you on a wild goose chase...



  • @snoofle said:

    I had an issue about 15 years ago where if our (C-motif) app was running, all the dialogs on another app turned to black-on-black making them (somewhat) more unusable than usual. Turns out that both apps were modifying the X-lib global colormap, and foregrounding either app would obliterate the GUI components of the other (until you switched back).

    At least that's intended and well-documented behavior.  The same thing would happen when you palette-swapped on Win3.1. It was typical among games, but often unnoticed because games would go full-screen before swapping the palette.

    Come to think of it I think this behavior is actually part of the VGA standard too.



  • @snoofle said:

    @djork said:

    @purge said:

    The mere presence of your app makes the other app crash.  Sounds like there are some serious issues to be resolved.

    Actually no.  We've found the issue for them, and it involves application A causing application B to crash when using feature X.

    We are the developers of an unrelated application C. 

    Bonus points if you got them to take you out for drinks after they sent you on a wild goose chase...

    Or billed them.

    Or took THEM out for drinks and then bailed on the bill. 



  • @djork said:

    @purge said:

    The mere presence of your app makes the other app crash.  Sounds like there are some serious issues to be resolved.

    Actually no.  We've found the issue for them, and it involves application A causing application B to crash when using feature X.

    We are the developers of an unrelated application C. 

    What you should have said to them was:

    "We have isolated the cause of this problem and determined that it has nothing to do with us. This took us 6 hours of analysis, valued at $150 per hour. Here's an invoice for work done outside the limits of our normal support. We'll tell you the solution when we get the cheque."



  • @teqman said:

    The same thing would happen when you palette-swapped on Win3.1. It was typical among games, but often unnoticed because games would go full-screen before swapping the palette.

    Come to think of it I think this behavior is actually part of the VGA standard too.

    Ah, you might know. Windows 3.1 set aside a block of 20 colours for itself -- the "standard" 16, plus four more truly useless colours on top (mint green, peach, lilac and mid-grey, but no orange). It's been a long time, but was it possible to ever re-allocate those, or were they reserved to keep the system interface usable?

    There's a toy called LavaLamp for Win 3.1 which, if you ask it to remap all 256 palette entries, it crashes, both under Win 3.1 and in 2000. You can only set it to remap 236 or lower. My best guess has been that Windows won't let you change the reserved colour set, but I don't know.

    The Real WTF is that the Cirrus Logic GL5426/8 drivers allocated a different standard 16 to Microsoft's drivers, breaking all my icon editors (for some reason they draw black if you alter the reserved colour set to something else). I rang up Microsoft and asked if they had any generic drivers for 1024x768 resolution, but an angry Scots woman insisted that no video drivers shipped with Windows. Of course they did, but just not 1024x768 :/


Log in to reply