forums.silverfrost.com

[sparge]

Trying to open any existing FTN95 project file in Plato generates a MS Visual C++ runtime library error, as below.

Buffer overrun detected!

Program: C:\apps\salford software\ftn95\plato3.exe

A buffer overrun has been detected which has corrupted the programs internal state. The program cannot safely continue execution and must now be terminated.

[PaulLaidler]

Try deleting (or renaming) the project .ini file.

[sparge]

Doesn't work

I've now established the same thing happens with a newly-created project. Also, the dependencies are handled incorrectly. I've filed a bug report.

If I uninstall, do I uninstall back to 4.91 or back to nothing there at all?

[PaulLaidler]

Off hand I don't know what you will go back to.

If you still have access to an earlier Plato, how about installing it somewhere else and copying it across. Everything else (other than Plato) should be OK. Assuming that others have the same problem with the new Plato, I will release a patch very soon. At the moment I cannot reproduce the problem so I am assuming that it is just the build for the release that has gone wrong.

[PaulLaidler]

Andy

To speed things up, if you wish you could send me a sample project with the kind of dependencies that cause the failure. I can then test to see if there is a new bug in the Plato dependency checking process.

[sparge]

Paul,

I've renamed the v5.0 Plato and copied across my previous version, and it is working OK so far - no buffer overruns, and dependencies handled OK.

I've sent a simple project illustrating the mishandled dependencies.

Andy

[PaulLaidler]

This turned out to be a significant regression with the effect that the latest release of Plato cannot be used for projects.

Please use an older version of Plato for the next few days until we can upload a patch or a new release.

[sparge]

Paul,

Hi. Bad news about the regression.

I upgraded on my home PC (also previously at v4.91) to v5.0 as well, and the buffer overrun problem did not manifest. That triggered me to do some more investigation on my work PC, to try and provide more information to help you reproduce the problem - I don't want to have to continue to use an old version of Plato when you fix the regression!

Initially, I couldn't reproduce the problem either , but now I can , and the process of figuring out what had changed gave me the necessary clue to the villain of the piece: the Plato ini file. After the problem became apparent, the way I went about using the previous version of Plato was to rename the v5.0 version, and use the previous version, un-renamed. The result is that if the new, renamed, version is used, it has no ini file to refer to, first time, and has to generate a very rudimentary version afresh upon first use - and under these circumstances the problem does not occur. Once I twigged this , I renamed the old version instead and "un-renamed" the v5.0 version, and was able to reproduce the problem.

At this point, I used Dependency Walker to log the process up to the point where the buffer overrun message is posted.

I will email you a copy of my plato3.ini file and the dependency walker log file. The first ought to allow you to reproduce the problem; the second ought to help you diagnose it if not.

Good luck!

Andy

[PaulLaidler]

Andy

Thanks for the offer but I have already identified and fixed the regression.

Regards

Paul

[sparge]

Apologies, Paul - I thought the dependency-handling problem was the regression, and the buffer overrun thing was something else unrelated.

[MERO]

Hello
I just updated to V 5.0
I get the same problem as described by by Sparge using Plato.

Any help ?

Klaus

[PaulLaidler]

Version 5.01 is about to be uploaded.