Regression bug: Application fails to start with v4.0.9.2 (worked with v4.0.9.0)Solved

We have a 64-bit VC14-complied application that no longer runs (hangs before program entry) when using TA 4.0.9.2.

Comparing the 4.0.9.0 and 4.0.9.2 DLLs with Dependency Walker reveals no obvious differences... I'm hoping it can be easily reproduced by the LimeLM team.

Thanks,Nick

Hey Nick,

We can't reproduce this here. Make sure the dll version for both your x86 and x64 version is at least 4.0.9.2. Version 4.0.9.0 had a crashing bug in the x86 version.

If you're still having problems we need more information (because we can't reproduce this here based on the information provided):

1. The version of windows you're using.

2. The language you're using (C/C++ presumably)?.

3. (optional, but useful): Steps to reproduce this. Either example code or something.

4. The exact function you're calling with parameters that casues the "hang" (is it a crash or a hang?)

Checking in with the same problem (?)I have a user who couldn't activate pre-9 (adapter issue), got a crash with 4.0.9 and now gets a spinning progress bar on 4.0.9.2. I can't replicate the issue either, and am trying to get some more system information from the user.

Any more information would be ideal. We're going to try hard to reproduce this, but we've had no luck so far. Can you also answer questions 1 through 4 above. Also, are you using the TurboAcativate Wizard? Is that what's hanging?

My user has tried via our app and also running TA directly with the same result (no observable crash, just a hang). He is running W7 Home Premium 64 bit.

Ok, we'll dig into this. Nick, if you can also get back to me with more information that would be helpful.

Hi Wyatt,

I'm pretty sure it's a dependency problem with 4.0.9.2.

4.0.9.2 actually works on our development machines (which have various SDKs installed and I assume some unknown dependency 4.0.9.2 requires) but on our production machines, our software hangs (does not even reach program entry) - so something is missing.

What kind of info would help you solve this?

Cheers,Nick

Wyatt wrote:> Hey Nick,> > We can't reproduce this here. Make sure the dll version for both your x86> and x64 version is at least 4.0.9.2. Version 4.0.9.0 had a crashing bug in> the x86 version.> > If you're still having problems we need more information (because we can't> reproduce this here based on the information provided):> > 1. The version of windows you're using.> > 2. The language you're using (C/C++ presumably)?.> > 3. (optional, but useful): Steps to reproduce this. Either example code or> something.> > 4. The exact function you're calling with parameters that casues the> "hang" (is it a crash or a hang?)

1. Windows 7 Professional x642. C++ compiled with VC14 (Visual Studio 2015 Update 3)3. Just running our software which links to the 4.0.9.2 TA DLL4. Actually, the software does not even start - it just hangs before program entry, which leads me to believe it is a dependency problem.

Nick

It sounds like a dependency problem of some sort (a depend your not behaving correctly -- not that it's missing).

We're going to attempt to reproduce this tonight and early tomorrow. And if we can reproduce this we'll work on a fix ASAP.

We've could reproduce the bug on Windows 7 x64 and Windows Vista x64 using the x64 binaries of TurboActivate, TurboFloat, and TurboFloat Server. We've released 4.0.9.3 which fixes this bug. Get it here: https://wyday.com/limelm/api/

Sorry about the inconvenience.