Re: [HACKERS] Consistently catch errors from Python _New() functions

2017-11-18 Thread Tom Lane
Peter Eisentraut writes: > On 11/17/17 12:16, Tom Lane wrote: >> I'm confused by the places that change PLy_elog calls to pass NULL: >> >> -PLy_elog(ERROR, "could not create globals"); >> +PLy_elog(ERROR, NULL); >> >> How is that an

Re: [HACKERS] Consistently catch errors from Python _New() functions

2017-11-18 Thread Peter Eisentraut
On 11/17/17 12:16, Tom Lane wrote: > I'm confused by the places that change PLy_elog calls to pass NULL: > > - PLy_elog(ERROR, "could not create globals"); > + PLy_elog(ERROR, NULL); > > How is that an improvement? Otherwise it looks reasonable. If we pass NULL, then

Re: [HACKERS] Consistently catch errors from Python _New() functions

2017-11-17 Thread Tom Lane
Peter Eisentraut writes: > While reviewing some unrelated code, I noticed that we are handling > error conditions from Python API functions such as PyList_New() and > PyDict_New() in pretty random ways or not at all. Here is a patch to > fix that. This needs