Puzzling print-driver bug

[jmk_phd]

A week ago I rebuilt a year-old app for Windows using LC 9.6.9 on the assumption that this version of LC had corrected the bug involving default print drivers in recent versions of Windows 10/11. After code-signing it and the .msi installer, I found that all worked just fine (at least on my copy of Win10 via Mac Parallels). Great!

Then just the other day I did the same for another year-old app, but found in this case that the LC 9.6.9 standalone would not open! However, when I rebuilt it etc. using the LC 10 rc5, it opened and ran just fine.

I’m both puzzled and curious as to what it is that might cause one app to work with the presumed LC 9.6.9 fix, yet another to not. Has anyone identified anything specific to an app (e.g., inclusions) associated with vulnerability to the print-driver bug?

Frankly, I don’t care that much about those freeware apps, but I am about to market an app that must be reliable. (If the preview version of this app – which is available as an anonymous download – does not work reliably, no one is going to bother purchasing the full version.) I do know that the previous version had succumbed to the bug. (My first customer thankfully was savvy enough to follow my instruction to delete the default print drivers, but persons trying out the preview version are likely to just give up on it.)

I avoid employing rc versions of software – although I do understand the important contribution of those who do – but in this case waiting for the release of the final version of LC 10.0 will delay my marketing effort possibly by months.

What would you advise?

jeff k

[LiveCode_Panos]

Dear Jeff,

The fix to the original issue (i.e. crash on Windows 10/11 if some generic printer drivers are installed) is included in both LC 9.6.9 and 10 DP-5. So it is strange why you see different results in your other app in 9.6.9 vs 10 dp-5.

In my tests (with both 9.6.9 and 10 dp-5) I observed that although the app does no longer crash on startup - it might take longer to open. In some cases it took almost ~60 seconds to launch. So, I was wondering, could it be the case that in 9.6.9 the app does actually open, but it takes too long, so you thought that it did not open?

If you reliably still see this behavior (i.e. the app does not open in LC 9.6.9 but does so in 10 DP-5) then feel free to send us your stack to support@livecode.com and we'll investigate this issue further.

Kind regards,
Panos
--

[jacque]

Jeff, does the delay happen only on first launch of a new installation, or does it occur on every launch?

[jmk_phd]

Hi Panos and Jacque –

As far as I can tell, there was a glitch – I don’t know what or why – in the first build of this app using LC 9.6.9 that was responsible for its failure to open in Windows.

I have since rebuilt this Windows app several times using LC 9.6.9 : from copies of the stack that last had been saved in 10dp5, in 9.6.9, and unchanged from whatever version of LC was current back in August 2021. All standalones open and run just fine, so presumably the problem was not with the scripting itself.

Inasmuch as someone *did* encounter the print-driver bug with my far more important commercial app (built a few months before 9.6.9 was released), I was primed to jump to the conclusion that this same bug must have been responsible here as well, without considering the possibility of a glitch in the build instead.

The lesson learned is that the problem is not always the most obvious target.

It is reassuring that the print-driver bug has largely been squashed with 9.6.9 and 10dp5, although the occasional long delay in launching mentioned by Panos – presumably only in certain versions of Windows – is still a bit troubling, inasmuch as it might dissuade some folks from considering a particular app fully functional.

Thanks so much to you both for your replies!

jeff k