On Sat, 7 Sep 2019, Markus Elfring wrote: > Hello, > > I have taken another look at a known script for the semantic patch language. > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/scripts/coccinelle/api/ptr_ret.cocci?id=1e3778cb223e861808ae0daccf353536e7573eed#n3 > > I got the impression that duplicate SmPL code can be reduced here. > So I tried the following approach out. > > … > @depends on patch@ > expression ptr; > @@ > ( > ( > - if (IS_ERR(ptr)) return PTR_ERR(ptr); else return 0; > | > - if (IS_ERR(ptr)) return PTR_ERR(ptr); return 0; > ) > + return PTR_ERR_OR_ZERO(ptr); > | > - (IS_ERR(ptr) ? PTR_ERR(ptr) : 0) > + PTR_ERR_OR_ZERO(ptr) > ) > … > > > Unfortunately, I got the following information then for a test transformation. > > elfring@Sonne:~/Projekte/Linux/next-patched> spatch -D patch scripts/coccinelle/api/ptr_ret.cocci drivers/spi/spi-gpio.c > … > 29: no available token to attach to > > > It seems that the Coccinelle software “1.0.7-00218-gf284bf36” does not like > the addition of the shown return statement after a nested SmPL disjunction. > But the following SmPL code variant seems to work as expected. > > > … > @depends on patch@ > expression ptr; > @@ > ( > - if (IS_ERR(ptr)) return PTR_ERR(ptr); else return 0; > + return PTR_ERR_OR_ZERO(ptr); > | > - if (IS_ERR(ptr)) return PTR_ERR(ptr); return 0; > + return PTR_ERR_OR_ZERO(ptr); > | > - (IS_ERR(ptr) ? PTR_ERR(ptr) : 0) > + PTR_ERR_OR_ZERO(ptr) > ) > … > > > How do you think about to reduce subsequent SmPL rules also according to > a possible recombination of affected implementation details? There is not going to be any change with respect to this issue. It's fine when replacing one statement by another, but introduces complexity when removing something more complex. And there's not point to have something that works in only one special case. julia
