“Compiler Construction” by Niklaus Wirth (2014) [pdf] | Hacker News Reader

“Compiler Construction” by Niklaus Wirth (2014) [pdf]

inf.ethz.ch

127 pointsnetworked10 comments

I remember Wirth's approach of preferring recursive-descent over the bottom-up LR, LALR, SLR, etc., was quite controversial at the time, since prevailing wisdom was that bottom-up parser-generators could parse more complex grammars and were widely believed to be more efficient. Now it seems things have turned in the opposite direction, with both large production-quality compilers and simple educational "toy" ones using some variant of recursive-descent. When GCC switched from a Bison-generated parser to recursive-descent, it actually made it faster(https://gcc.gnu.org/wiki/New_C_Parser ). The fact that LL languages are theoretically a subset of LR has not stopped recursive-descent from being used to successfully parse languages like C++ whose grammar is actually Turing-complete.

The Digital Mars C++ compiler used recursive descent since 1987 or so. It's certainly practical.

Slightly off topic, but as one of the only people in the world who's written a C++ compiler by themselves, how many manhours would you estimate it would take to update a C++03 compiler to fully support C++14? Or how many years; I think I remember reading you say somewhere that DMC++ took 10 years to code.

I think I'm the only person who's written a complete C++ compiler by himself - preprocessor, parser, semantic analyzer, optimizer, and code generation. I also mean a C++98 compiler, because that's pretty much where I stopped.

I don't know how much is involved in going from that to C++14, as I haven't done that.

10 man years is a good estimate for creating a C++ compiler, not including the C part nor the optimizer/codegen.

It's great to see this available and updated. It was one of my favorite compiler books growing up because it's so small and straight-forward. I feel it's a much better starting point for someone actually interested in building a compiler than say, the Dragon book.

Even more immediate and hands-on is Abdulaziz Ghuloum's "An Incremental Approach to Compiler Construction" paper: http://scheme2006.cs.uchicago.edu/11-ghuloum.pdf (which chides Wirth for oversimplifying)

It's a classic. Also helps you understand and start to play with a well-documented OS and compiler suite coded in a safe, GC language. Good start on developments where countering subversion of toolchain is a requirement. Wirth's stuff is one of least likely to be subverted & easiest to spot such things.

I didn't realise this book existed. My university was really into Pascal and Modula-2 and had a second or third year course called Compiler Construction. We used the Dragon book (which I really enjoyed), and recursive descent. I'm not sure if we actually used lex or yacc.

The 2014 version is available here:

https://www.inf.ethz.ch/personal/wirth/CompilerConstruction/...

Thanks, we changed to that from http://www.ethoberon.ethz.ch/WirthPubl/CBEAll.pdf.

I actually prefer the '05 edition, as it targets a RISC ISA. (I also reach for my 2nd ed of The Art of Electronics more than my 3rd, take that as you will). Thanks for linking that out, a classic read along with Crenshaw.

Reply on news.ycombinator.com