Microservices everywhere. Those are the times we live in now. Everyone seems to be splitting monolithic web applications into smaller chunks. And that's fine. However, setting up the local development environment for this can be sometimes a bit cumbersome.

At this point we should have a reasonable picture of how flang parses and generates semantic information. So now it is time to explore with more detail what is actually synthesized and how it can be used later in the compiler. In this chapter we are going to see the symbol table.

I’ve been implementing an algorithm that works on a graph. That algorithm needs to create and destroy lots of nodes and edges to keep track of the algorithm state correctly. The algorithm also needs to be fast in order to be competitive against a similar algorithm that uses sets of stacks instead of graphs. Profiles show that memory allocations are impacting negatively the performance, so maybe a memory pool can help.

In chapter 6 we saw conditional branches and we ended commenting that they can be used to implement higher control constructs. In this chapter we will see a few of them.

In the previous installment of this series we saw how flang parses the statements using an LR(1) algorithm. As the parser recognized the parts of the statements it invokes semantic actions. Today we’re going to talk more about them.