On Jan 19, 2011, at 5:36 PM, Kevin Grittner wrote: > "A.M." <agentm@xxxxxxxxxxxxxxxxxxxxx> wrote: > >> Most PLs include some session-specific storage. In PL/Perl, it is >> %_SHARED. Setting a flag there should do the trick. If you are >> using a PL which does not have such a notion (like plpgsql), you >> can add a call in your triggers to a function written in a PL >> which does support this. Alternatively, a C function which >> sets/checks a global flag would work as well. > > I thought it might come to that. I'm comfortable writing C > functions, and we're not using any languages so far besides C, SQL, > and plpgsql, so I'd probably use C. If I'm going that far, though, > I'd be rather inclined to implement a TG_DEPTH variable (as being > easier for us to use) and offer it to the community in case there's > anyone else who would find this useful. If that turns out to be > harder than I think, I'll fall back to what you outlined here. If you do implement TG_DEPTH, I am curious as to what the difference between TG_DEPTH==34 and TG_DEPTH==35 could mean. I think it might cause poor coding practice in making decisions based off assumed trigger order execution. Since you only care to distinguish between depth 1 and depth 2 (and not beyond), could you elaborate on a use case where further trigger "depth" information may be useful? Cheers, M -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
