On Mon, Apr 9, 2012 at 1:23 PM, Julia Lawall <[email protected]> wrote: > On Mon, 9 Apr 2012, Jim Cromie wrote: > >> On Mon, Apr 9, 2012 at 6:43 AM, Julia Lawall <[email protected]> wrote: >>> >>> For the whole kernel, I get over 32,000 potential added calls. Perhaps >>> there is some way to better constrain the cases where it should be used? >>> >>> julia >> >> >> OOH, THANK YOU for that script. >> Now that I see it, I understand it - at least a little bit. >> You are a ninja !! >> >> 32k is *alot* >> and I see youve already constrained it to static defs. >> >> Devs sometimes add comments about keeping A,B in sync if you change... >> but the form of comments surely varies, and IIUC, cocci cant look at >> comments. > > > No, it doesn't look at comments. > > >> I'll try it out, inspect the results, and see if I can find more >> constraints. > > > Actually, the previous results were all backwards. All the constraints > referred to arrays that were declared afterwards. I've attached two new > versions. Tables.cocci collects constraints to all previous arrays of the > same size. Last_tables.cocci puts a constraint only for the most recent > table - if all of the constraints are satisfied it should amount to the same > thing. Last_tables generates less code, but the result is more fragile > because the most recent table is not necessarily the one you want. > > julia
I'll start with last-tables, see how it looks. Do you recall how many patches it created ? Less is better for now.. What does it do with table A fns ... table B more fns table C ? + BUILD_BUG_DECL(AC, ...); + BUILD_BUG_DECL(BC, ...); would be close to ideal. _______________________________________________ Cocci mailing list [email protected] http://lists.diku.dk/mailman/listinfo/cocci (Web access from inside DIKUs LAN only)
