From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-10-14 10:23:00
At 08:42 PM 10/13/2002, Joel de Guzman wrote:
>----- Original Message -----
>From: "Carl Daniel" <cpdaniel_at_[hidden]>
>> More review to come, but in the meantime, a question: have the Spirit
>> developers attempted to measure the performance of Spirit-generated
>> against other parser generators (e.g. YACC, Bison, ANTLR)? Similarly,
>> have comparisons of generatd code size been done?
>There were informal benchmarks here and there, not only for code size and
>runtime performance but for compile time as well. It should be the right
>time to formally compile them and produce a report.
While space/speed comparisons are always interesting, there are other
characteristics which often drive the decision as to which parser is best
for an application. For example,
* Usefulness and malleability of error messages, particularly when
dying on the first error encountered isn't an option.
* Ability to make small changes to the parsed language without
disastrous rewrites of existing parser code and/or tables.
* Can easily parse local contradictions (always AB,
except in certain expressions its BA).
My past experience has been that recursive decent parsers are very strong
in those kinds of "quality" areas. I'd like to get some indication of how
Spirit does with such quality related issues.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk