> On 29 May 2019, at 15:03, Tom Lane <t...@sss.pgh.pa.us> wrote: > > Pursuant to today's discussion at PGCon about code coverage, I went > nosing into some of the particularly under-covered subdirectories > in our tree,
On a similar, but much less important/interesting note. I fat-fingered when compiling isolationtester on the plane over here and happened to compile src/test/examples, and in there testlo.c and testlo64.c has two dead functions for which the callsites have been commented out since the Postgres95 import (and now cause a warning). Is there any (historic?) reason to keep that code? It also seems kind of broken as it doesn’t really handle the open() call failure very well. cheers ./daniel
