Re: pg_dumpall bug & feature request - Mailing list pgsql-bugs

From Ivan
Subject Re: pg_dumpall bug & feature request
Date
Msg-id 153662571.20051103173305@mail.ru
Whole thread Raw
In response to Re: pg_dumpall bug & feature request  (Alvaro Herrera <alvherre@commandprompt.com>)
Responses Re: pg_dumpall bug & feature request
List pgsql-bugs
Hello Alvaro,

Thursday, November 3, 2005, 3:28:41 PM, you wrote:

AH> Ivan wrote:

AH> Hi,

>> Perhaps you missed my previous message.
>> pg_dumpall has not -f command line option
>> to specify output file - it always send output
>> to standart output. But there is an issue with
>> it in Windows (I described it earlier) - function's
>> line endings are changed from 0D 0A to 0D 0D 0A.

AH> Hmm, what do you need that for?  Why don't you directly restore the
AH> binary dump to a database by using option -d?

I don't need to restore binary dump. I need to make a dump of the
whole database cluster to plain sql file (what pg_dumpall does).
So I have to redirect pg_dumpall's output to a file
pg_dump [options] > cluster_dump.sql
When I do it on Windows, it creates cluster_dump.sql file
BUT in that file all sql and plpgsql function bodies are corrupted! -
each line in the body ends with 0D 0D 0A sequence instead of 0D 0A -
therefor functions after restoring looks like:

CREATE OR REPLACE FUNCTION "foo"()
RETURNS bar AS
$$

  select * from "bar"

  where  bla-bla

  order by bla-bla;

$$
  LANGUAGE 'sql' VOLATILE;

instead of

CREATE OR REPLACE FUNCTION "foo"()
RETURNS bar AS
$$
  select * from "bar"
  where  bla-bla
  order by bla-bla;
$$
  LANGUAGE 'sql' VOLATILE;

 - extra blank lines are added.

So it would be great if the output redirection ">" will be workaround
for example the same way like in pg_dump it is done - via -f file
option.

Thank you for support.

--
Best regards,
 Ivan                            mailto:Ivan-Sun1@mail.ru

pgsql-bugs by date:

Previous
From: "Jolly Chen"
Date:
Subject: BUG #2017: column labels ignored on selects from views
Next
From: Tom Lane
Date:
Subject: Re: Failed CLUSTER seems to leave files behind