Why is ecpg segfaulting on buildfarm member "clownfish"?

Has anyone looked into $SUBJECT? I just today realized that the
ECPG-Check failures on that machine are not the run-of-the-mill
small difference in the expected results. Rather, most of the
tests are actually dumping core on the client side:

testing connect/test1.pgc ... skipped
testing connect/test2.pgc ... Segmentation Fault - core dumped
FAILED (log)
testing connect/test3.pgc ... Segmentation Fault - core dumped
FAILED (log)
testing connect/test4.pgc ... ok
testing connect/test5.pgc ... ok
testing compat_informix/charfuncs.pgc ... ok
testing compat_informix/dec_test.pgc ... ok
testing compat_informix/rfmtdate.pgc ... ok
testing compat_informix/rfmtlong.pgc ... ok
testing compat_informix/rnull.pgc ... Segmentation Fault - core dumped
FAILED (log, output)
testing compat_informix/test_informix.pgc ... Segmentation Fault - core dumped
FAILED (log, output)
testing compat_informix/test_informix2.pgc ... Segmentation Fault - core dumped
FAILED (log, output)
testing preproc/comment.pgc ... ok
testing preproc/define.pgc ... Segmentation Fault - core dumped
FAILED (log, output)
etc etc

BTW, this is a perfect example of why it's not a good idea to allow
minor regression failures to go unfixed --- people become desensitized.
I know I've been completely ignoring ECPG-Check buildfarm results
for awhile now.

regards, tom lane