A std::io::Read-based parser. Also, ran cargo format.
#3
Conversation
This change is straightforward because the DIMACS format is ASCII-only.
Panics on IO errors Why? Because the alternatives are: * `collect` the first IO error -- requiring us to pull the whole string into memory * catch those IO errors deep in the lexer, and bubble them up as a special kind of `ParseError`. Since we'll generally be reading form files, read errors won't typically happen, so I figure its better to just panic, since a use wouldn't want to recover any ways.
1 similar comment
Didn't know that the |
I'm not using it yet, but I plan to write a set of teaching SAT solvers on top of it. |
Good to know then we might want to improve it and also provide a streaming interface for clauses using the https://crates.io/crates/ipasir crate. |
|
Thanks for the PR. I merge this now. |
The reason for this is that it allows parsing to begin without reading the whole file into memory.
At some point it might be nice to add a "clause stream" output as well---then you could incrementally read clauses and assimilate them directly into the SAT solver data structures.
I chose to panic on IO error during the read, because I don't think that users would want to catch them, and panicking is much simpler.
Also, I ran cargo format. Sorry about the large diff, but my editor handles tabs poorly.