On 03/26/2014 09:27 AM, Heikki Linnakangas wrote:
> On 03/26/2014 03:15 PM, Andrew Dunstan wrote:
>>
>> On 03/26/2014 07:50 AM, Heikki Linnakangas wrote:
>>> Add -f/--follow option to pg_xlogdump.
>>>
>>> This is useful for seeing what WAL records are inserted in
>>> real-time, by
>>> pointing pg_xlogdump to a live server.
>>>
>>> Branch
>>> ------
>>> master
>>>
>>> Details
>>> -------
>>> http://git.postgresql.org/pg/commitdiff/ce9bb92f8fb8d25cf00ec939797ffdb5930fb792
>>>
>>>
>>> Modified Files
>>> --------------
>>> contrib/pg_xlogdump/pg_xlogdump.c | 31
>>> ++++++++++++++++++++++++++++---
>>> doc/src/sgml/pg_xlogdump.sgml | 11 +++++++++++
>>> 2 files changed, 39 insertions(+), 3 deletions(-)
>>>
>>
>>
>> Seems to have upset Windows (see
>> <http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=currawong&dt=2014-03-26%2012%3A30%3A44>).
>>
>>
>> Shouldn't you be using pg_usleep() instead of sleep()?
>
> Apparently. Fixed that way.
>
> I'm surprised that pg_upgrade doesn't have the same problem - there's
> a plain sleep(1) call in contrib/pg_upgrade/exec.c. What's the difference?
pg_upgrade.h has:
#ifdef WIN32
#define sleep(x) Sleep(x * 1000)
#endif
Which is another way of doing it, I guess. I'm not sure why the usual
way wasn't adopted there.
cheers
andrew