Re: WAL replay bugs

Hello, The new patch looks good for me.

The usage of this is a bit irregular as a (extension) module but
it is the nature of this 'contrib'. The rearranged page type
detection logic seems neater and keeps to work as intended. This
logic will be changed after the new page type detection scheme
becomes ready by the another patch.

I have some additional comments, which should be the last
ones. All of the comments are about test.sh.

- A question mark seems missing from the end of the message "Has
build been done with -DBUFFER_CAPTURE included in CFLAGS" in
test.sh.

- "make check" runs "regress --use-existing" but IMHO the make
target which is expected to do that is installcheck. I had
fooled to convince that it should run the postgres which is
built dedicatedly:(

- postgres processes are left running after
test_default(custom).sh has stopped halfway. This can be fixed
with the attached patch, but, to be honest, this seems too
much. I'll follow your decision whether or not to do this.
(bufcapt_test_sh_20140710.patch)

- test_default.sh is not only an example script which will run
while utilize this facility, but the test script for this
facility itself.

So I think it would be better be added some queries so that all
possible page types available for the default testing. What do
you think about the attached patch? (hash index is unlogged
but I dared to put it for clarity.)

(bufcapt_test_default_sh_20140710.patch)

regards,

--
Kyotaro Horiguchi
NTT Open Source Software Center