Re: [HACKERS] new warning

2011-11-09 Thread Greg Jaskiewicz
On 9 Nov 2011, at 16:16, Robert Haas wrote: > On Wed, Nov 9, 2011 at 11:12 AM, Kevin Grittner > wrote: >> This commit adds a new warning on my machine: >> >> http://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=d326d9e8ea1d690cf6d968000efaa5121206d231 >> >> copy.c: In function *Do

Re: [HACKERS] new warning

2011-11-09 Thread Robert Haas
On Wed, Nov 9, 2011 at 11:12 AM, Kevin Grittner wrote: > This commit adds a new warning on my machine: > > http://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=d326d9e8ea1d690cf6d968000efaa5121206d231 > > copy.c: In function *DoCopy*: > copy.c:1861:14: warning: *bufferedTuples* may be

[HACKERS] new warning

2011-11-09 Thread Kevin Grittner
This commit adds a new warning on my machine: http://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=d326d9e8ea1d690cf6d968000efaa5121206d231 copy.c: In function *DoCopy*: copy.c:1861:14: warning: *bufferedTuples* may be used uninitialized in this function -Kevin -- Sent via pgsq

Re: [HACKERS] new warning message

2008-02-27 Thread Tom Lane
Jeff Davis <[EMAIL PROTECTED]> writes: > Would it be reasonable to throw a warning if you revoke a privilege from > some role, and that role inherits the privilege from some other role (or > PUBLIC)? This has been suggested and rejected before --- the consensus is it'd be too noisy. Possibly the

[HACKERS] new warning message

2008-02-27 Thread Jeff Davis
On IRC today someone brought up a problem in which users were still able to connect to a database after a "REVOKE CONNECT ... FROM theuser". The reason theuser is still able to connect is because PUBLIC still has privileges to connect by default (AndrewSN was the one who answered this). Would it b

Re: [HACKERS] New warning in pg_dump

2004-08-23 Thread Tom Lane
Philip Warner <[EMAIL PROTECTED]> writes: > At 01:27 PM 24/08/2004, Tom Lane wrote: >> I prefer declaring it as unsigned, which means you drop the >> \377 end instead ... > ...I've used explicit values (128) since '\200' is -127. Actually I'd go with "(unsigned char) '\200'". There's a bunch of

Re: [HACKERS] New warning in pg_dump

2004-08-23 Thread Philip Warner
At 01:27 PM 24/08/2004, Tom Lane wrote: I prefer declaring it as unsigned, which means you drop the \377 end instead ... No problem. Philip Warner| __---_ Albatross Consulting Pty. Ltd. |/ - \

Re: [HACKERS] New warning in pg_dump

2004-08-23 Thread Tom Lane
Philip Warner <[EMAIL PROTECTED]> writes: > At 01:58 AM 24/08/2004, Alvaro Herrera wrote: >> static int _isIdentChar(unsigned char c) > I think the correct thing to do is to leave it as (signed) char, and remove > the comparison to \200 = -127. No, that isn't the right thing, because not all pla

Re: [HACKERS] New warning in pg_dump

2004-08-23 Thread Philip Warner
At 01:27 PM 24/08/2004, Tom Lane wrote: I prefer declaring it as unsigned, which means you drop the \377 end instead ... ...I've used explicit values (128) since '\200' is -127. Philip Warner| __---_ Albatr

Re: [HACKERS] New warning in pg_dump

2004-08-23 Thread Philip Warner
At 01:58 AM 24/08/2004, Alvaro Herrera wrote: static int _isIdentChar(unsigned char c) I think the correct thing to do is to leave it as (signed) char, and remove the comparison to \200 = -127. All chars will be >= -127. I will fix this in the next release. --

[HACKERS] New warning in pg_dump

2004-08-23 Thread Alvaro Herrera
Hackers, I'm seeing this: gcc -O2 -fno-strict-aliasing -g -Wall -Wmissing-prototypes -Wmissing-declarations -I/home/alvherre/CVS/pgsql/source/00orig/src/interfaces/libpq -I../../../src/include -I/home/alvherre/CVS/pgsql/source/00orig/src/include -D_GNU_SOURCE -DFRONTEND -c -o pg_backup_db.o