Was FORTRAN buggy 4324

Because a debugging session of a device driver that you just wrote doesn't need a test for every bug. We would have never shipped anything if we had to write a test for every bug we found during a major development project. Code-reading bugs might not have a scenario that could exercise it from the "outside".

In our development group, this software didn't exist. The development project was make it exist. You don't write a test for every bug you find in this situation.

You keep buttuming that a user mode test can exercise bugs in the monitor. This isn't often the case.

You're confusing maintenance and enhancements with development. I suspec this is because you don't do OS development as a primary job but probably do tweaks.

I'm almost saying it was impossible to do. For example, take TW's RP20 project report; there is no way to write user mode tests for each of those. I suppose what we did was put your concept of tests into the monitor code. But that was simply part of the programming.

hmm...is this difference a DEC-bias thingie? I buttumed that everybody did this kind of stuff; it's more natural than breathing.

I am talking about the coding that has to exist underneath all of those virtuosities. That is the development I'm talking about. A virtual machine is an app. Period.

No. You cannot write tests exercising exec level bugs that run at user mode. YOu can, with great difficulty and tons of time, maybe, write something that may exercise the code. But this is a waste of time if you have 500 bugs waiting for you before ship.

BAH