Re: Unstable timestamp binary representation? - Mailing list pgsql-hackers

From Shachar Shemesh
Subject Re: Unstable timestamp binary representation?
Date
Msg-id 423B33D3.3060306@shemesh.biz
Whole thread Raw
In response to Re: Unstable timestamp binary representation?  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-hackers
Tom Lane wrote:

>Shachar Shemesh <psql@shemesh.biz> writes:
>  
>
>>In other words, it seems that I, as a client, needs to guess whether 
>>postgres was compiled with or without "HAVE_INT64_TIMESTAMP".
>>    
>>
>
>No, you need to inquire of the value of the "integer_datetimes"
>parameter.  (At least as of 8.0, this is provided "for free" during
>connection startup, so you don't even need an extra network round
>trip to find it out.)
>  
>
Hmm. So I need to call "|PQparameterStatus" with "|integer_datetimes". 
If I get "TRUE", treat as integer, if I get "FALSE", treat as double. 
What happens if I get NULL? The docs seem to suggest that this is what 
will happen for 7.4 back end, and the FALSE is the default. My 
experience shows, however, that at least 7.4 for Debian Sid should be true.

>No, it checks.  See pg_control.
>  
>
Not what I have seen from the code, but I will take your word for it.

>            regards, tom lane
>  
>         Shachar

-- 
Shachar Shemesh
Lingnu Open Source Consulting ltd.
Have you backed up today's work? http://www.lingnu.com/backup.html



pgsql-hackers by date:

Previous
From: "Dann Corbit"
Date:
Subject: Re: "they only drink coffee at dec"
Next
From: "Dave Page"
Date:
Subject: Re: [pgsql-www] "they only drink coffee at dec"