'Writing a C Compiler' is a book (2022)
norasandler.com
norasandler.com
However, looking at the table of contents [0], she seems to spend a lot of the book on types and pointers. That's cool, since the compiler books that I've seen don't cover languages with pointers and pointer arithmetic (and the analyses that come with that), or representations of types.
C's selling point is performance and portability and this book covers neither
https://github.com/rui314/chibicc
> Each commit of this project corresponds to a section of the book. For this purpose, not only the final state of the project but each commit was carefully written with readability in mind. Readers should be able to learn how a C language feature can be implemented just by reading one or a few commits of this project.
I don't understand this criticism. It seems reasonable for the end-result compiler to support only a 'significant subset of C', rather than total compliance with the standard. It also seems reasonable for the performance of its generated code to be nowhere near that of the major optimising C compilers.
How could it possibly be otherwise? A serious modern C compiler is the result of millions of dollars of specialised software engineering.
> C's selling point is performance and portability and this book covers neither
I've not read the book, but it apparently does cover optimisation.