On Tue, 23 Feb 2010, SF Markus Elfring wrote: > > The patterns should be as close to C as possible. Thus, in general syntax > > extensions are unacceptable. > > That is a pity. - I imagine that more programming and scripting languages > could > benefit from the semantic patch language.
Oops, sorry, I didn't mean to preclude extending Coccinelle to other programming languages. That is just a matter of time, which we don't seem to hav bery much of :). What I meant to preclude was adding a lot of non-concrete-syntax-like constructs in the pattern part, to express eg the type of a function call. > > Ideally, the user should be able to do everything he wants from knowing > > the patch format and from an example that shows how to declare some > > metavariables. > > I like this goal, too. > > > > I realize that the current state of the language might not meet up to > > that ideal :) > > About which open issues do you think here? The management of inheritance of metavariables is clearly rather subtle. There are some subtleties to the "when" construct as well, and the difference between <... ...> and <+... ...+> does not always seem to be obvious. > > It would certainly be possible to make a language that has a lot more > > kinds of different decorations, but that is not the goal of this > > particular project, and that might not meet the needs of quite the same > > user community. > > I hope that current limitations can be removed in the future. The support for > additional features will result in more work and efforts that you would like > to > avoid at the moment, wouldn't you? Some of it is the effort required. But some of it is also the desire to stick within a particular language paradigm, rather than making something that tries to do everything and solve all problems for everyone. julia _______________________________________________ Cocci mailing list [email protected] http://lists.diku.dk/mailman/listinfo/cocci (Web access from inside DIKUs LAN only)
