|Incremental compilation email@example.com (1992-01-23)|
|Re: Incremental compilation firstname.lastname@example.org (1992-01-23)|
|Re: Incremental compilation email@example.com (1992-01-23)|
|Re: Incremental compilation firstname.lastname@example.org (1992-01-24)|
|Incremental Compilation email@example.com (Alexander Rozenman) (1999-11-02)|
|Re: Incremental Compilation firstname.lastname@example.org (Matthew Economou) (1999-11-03)|
|Re: Incremental Compilation maratb@CS.Berkeley.EDU (Marat Boshernitsan) (1999-11-05)|
|[2 later articles]|
|From:||email@example.com (Steve Boswell)|
|Organization:||Free Software Foundation|
|Date:||23 Jan 92 08:19:49 GMT|
I'm looking for references, information, etc. on incremental compilation.
I don't mean makefiles -- the unit of compilation there is the file, and I
wanted something smaller, like the lexical token.
Here's a technique I was planning to try in a compiler. It would save the
entire token stream, syntax tree, semantic information, etc. for a
particular source file compilation. When the source file is modified and
recompiled, it would make a new token stream, do a token-wise diff on the
old and new streams, redo only the syntactic units that it had to, then
recheck semantics recursively.
Has anything like this been done/thought of/etc.? Doing this in my own
project is a long way off, but I was wondering if the time it took to do
the token-wise diff was longer than the time it would take to re-parse the
[It's been done a lot in systems with built-in editors, since the editor
can easily mark what's been changed. -John]
Return to the
Search the comp.compilers archives again.