Cannot run the *.EXE ?

[WSinc]

I occasionally get this message after a build:

"cannot run/open EXE"

This is just a nuisance, it usually requires me to do a CLEAN and rebuild.

It happens maybe 20 percent of the time.

Any clues as to why this occurs?

[Steven_L_Intel1]

I've never seen this before. If you look at the buildlog.htm, does this message show there and at what step of the build do you see it?

[WSinc]

Next time it happens, I can send you the buildlog.htm file.

I think it completes the build successfully, but some process is hanging up the EXE

so I can't start it when it's done.

I was wondering re the interface checking - doesn't it make more sense to do that in the LINKER phase?

That way it can make sure there aren't any omitted or duplicate entry points in the set of object files.

And the interface checking would then be independent of the languages used.

[anthonyrichards]

This can happen if you have not terminated a previous instance of the *.EXE, so the 'new' version of the exe cannot be built before the previous one is terminated and unlocked by the system to permit its deletion.

Sometimes an*.EXE can terminate abnormally leaving a stub locked by the system and requiring deletion using Task Manager.

[Steven_L_Intel1]

Sure, it makes sense to do interface checking in the linker phase. If you have control over all the compilers, the linker and the object language. It's very ambitious - the original design spec for the VAX/VMS linker and object language had this, but only a very small part of it ever got implemented.

Absent that, the compiler can do a good job within its constraints, and generally it does. Of course, as shown in this forum frequently, people are surprised by the additional diagnostics when compiling what they thought was correct code and in many cases would rather disable the checks rather than fix the source...