On 7/28/21 11:02 AM, Tom Lane wrote: > > Here I've got to disagree. We do need the form with a PQExpBuffer > argument, because there are some places where that isn't a pointer > to a PGconn's errorMessage. But the large majority of the calls > are "pqReportOOM(conn)", and I think having to write that as > "pqReportOOM(&conn->errorMessage)" is fairly ugly and perhaps > error-prone. > > I'm not wedded to the name "pqReportOOMBuffer" though --- maybe > there's some better name for that one? > >
Is it worth making the first one a macro? cheers andrew -- Andrew Dunstan EDB: https://www.enterprisedb.com
