Fabian Ruffy
Fabian Ruffy
>I suspect you could get a much bigger gain here by getting rid of the java-style Util::Enumerator (which does a bunch of dynamic allocation and generates much garbage to be...
Tracking a comment. We can maybe achieve this at some point in time. >Related to this, I've been working on code (in https://github.com/ChrisDodd/p4c/tree/cdodd-disablegc) to allow removing the garbage collector and...
The way SideEffectOrdering propagates source information causes it to duplicate the source information in some cases. Ideally, we want to point to either the left- or right-hand side of an...
This PR is meant to raise a discussion. We stumbled across this in #4357 where an incorrectly initialized cstring caused a bug. This PR changes the default initializer to use...
Improve the error handling in the test back ends. Instead of throwing exceptions everywhere, try to gracefully handle errors and bubble them upwards. Considering using a utility macro for that....
CompilerTarget, Target, and Context are all singleton, static mystery objects floating around the interpreter. They are initialized in a specific way that makes it hard to make changes or extend...
Bold refactor to replace boost::fmt with [fmt](https://github.com/fmtlib/fmt), which is closer to the C++20 version of fmt. This is one step to get rid of the majority of the boost dependency....
With #4275 we added the possibility to validate the P4Testgen options in an explicit, second step. We should expand this to also validate options on a per-target basis, e.g., for...
https://github.com/jafingerhut/p4-guide/tree/master/specifying-p4-architectures is a new way to define the structure of a P4 target which can simplify static analysis by defining the transitions and operations between individual program blocks. P4Testgen can...