This is a rewrite of the vm integration for uopz ... it improves many things ....
Little bit of justification: uopz invalidated runtime caches in order to function, and while it's the case that this is sufficient a good portion of the time, it means that there are some pretty strange edge cases when trying to use uopz to test legacy code. The edge cases could result in memory errors and faults, of the unavoidable variety. In implementing more of the business logic for handlers we have the opportunity to instrument the vm at the correct time, more of the time and correctly.
This will remain open for another day or so, while I complete testing on work test suites, and try to improve test coverage a bit more ... it will then be merged into master.
On a side note, this merge will drop support for 7.0, it simplifies the code a little in preparation for the additional complexity of 7.4, which we will soon have to support.
This is a rewrite of the vm integration for uopz ... it improves many things ....
Little bit of justification: uopz invalidated runtime caches in order to function, and while it's the case that this is sufficient a good portion of the time, it means that there are some pretty strange edge cases when trying to use uopz to test legacy code. The edge cases could result in memory errors and faults, of the unavoidable variety. In implementing more of the business logic for handlers we have the opportunity to instrument the vm at the correct time, more of the time and correctly.
This will remain open for another day or so, while I complete testing on work test suites, and try to improve test coverage a bit more ... it will then be merged into master.
On a side note, this merge will drop support for 7.0, it simplifies the code a little in preparation for the additional complexity of 7.4, which we will soon have to support.