Thread: pg_hba.conf additional comment re local line
The attached tiny patch will possibly help to avoid some confusion by Windows users about the "local" line in pg_hba.conf (and thus help reduce queries to us ;-) ). It also removes an essentially content-free suffix in 2 nearby comment lines. cheers andrew Index: src/backend/libpq/pg_hba.conf.sample =================================================================== RCS file: /projects/cvsroot/pgsql-server/src/backend/libpq/pg_hba.conf.sample,v retrieving revision 1.53 diff -c -w -r1.53 pg_hba.conf.sample *** src/backend/libpq/pg_hba.conf.sample 26 Aug 2004 16:50:05 -0000 1.53 --- src/backend/libpq/pg_hba.conf.sample 21 Sep 2004 16:41:30 -0000 *************** *** 60,67 **** # TYPE DATABASE USER CIDR-ADDRESS METHOD local all all @authmethod@ ! # IPv4-style local connections: host all all 127.0.0.1/32 @authmethod@ ! # IPv6-style local connections: host all all ::1/128 @authmethod@ --- 60,68 ---- # TYPE DATABASE USER CIDR-ADDRESS METHOD + # "local" is for Unix domain socket connections, and does not apply to Windows local all all @authmethod@ ! # IPv4 local connections: host all all 127.0.0.1/32 @authmethod@ ! # IPv6 local connections: host all all ::1/128 @authmethod@
Andrew Dunstan wrote: > The attached tiny patch will possibly help to avoid some confusion by > Windows users about the "local" line in pg_hba.conf (and thus help > reduce queries to us ;-) ). It also removes an essentially > content-free suffix in 2 nearby comment lines. Maybe initdb should just remove those lines on Windows. -- Peter Eisentraut http://developer.postgresql.org/~petere/
Peter Eisentraut wrote: >Andrew Dunstan wrote: > > >>The attached tiny patch will possibly help to avoid some confusion by >>Windows users about the "local" line in pg_hba.conf (and thus help >>reduce queries to us ;-) ). It also removes an essentially >>content-free suffix in 2 nearby comment lines. >> >> > >Maybe initdb should just remove those lines on Windows. > > > You are probably right. initdb's filtering capabilities are (by design) rather primitive - I didn't want to clutter it up with regex code - so it's a bit less easy than it might otherwise be. This patch was done as a "quick fix". If someone doesn't beat me to it I will try to look at initdb in the next week or so. cheers andrew
Andrew Dunstan <andrew@dunslane.net> writes:
> The attached tiny patch will possibly help to avoid some confusion by
> Windows users about the "local" line in pg_hba.conf (and thus help
> reduce queries to us ;-) ).
I was wondering if we could teach initdb to remove that line altogether
in Windows installations.
regards, tom lane
Tom Lane wrote:
>Andrew Dunstan <andrew@dunslane.net> writes:
>
>
>>The attached tiny patch will possibly help to avoid some confusion by
>>Windows users about the "local" line in pg_hba.conf (and thus help
>>reduce queries to us ;-) ).
>>
>>
>
>I was wondering if we could teach initdb to remove that line altogether
>in Windows installations.
>
>
>
See later email.
Of course it can be done. Probably the simplest way is a new small
routine called, say filter_token, which would remove lines containing a
given token:
static char **filter_token(char **lines, char *token);
Then we could have something like:
#ifdef WIN32 (orwhatever we are calling it now)
conflines = filter_token(conflines,"@remove-line-for-win32@");
#else
conflines = replace_token(conflines, "@remove-line-for-win32@","");
#endif
Incidentally, even Unix users frequently get confused about the "local"
line - one of the commonest newbie mistakes is to think it means
localhost. So marking it clearly as being for Unix domain sockets would
still be a Good Thing (tm).
cheers
andrew
> regards, tom lane
>
>---------------------------(end of broadcast)---------------------------
>TIP 7: don't forget to increase your free space map settings
>
>
>
The original of this seems to have gotten lost in the ether somewhere.
It might turn up some day ...
andrew
Tom Lane wrote:
>Andrew Dunstan <andrew@dunslane.net> writes:
>
>
>>The attached tiny patch will possibly help to avoid some confusion by
>>Windows users about the "local" line in pg_hba.conf (and thus help
>>reduce queries to us ;-) ).
>>
>>
>
>I was wondering if we could teach initdb to remove that line altogether
>in Windows installations.
>
>
>
>
I think this does what Tom and Peter suggested. I don't have a Windows
box to test it on, though.
cheers
andrew
Index: src/backend/libpq/pg_hba.conf.sample
===================================================================
RCS file: /projects/cvsroot/pgsql-server/src/backend/libpq/pg_hba.conf.sample,v
retrieving revision 1.53
diff -c -w -r1.53 pg_hba.conf.sample
*** src/backend/libpq/pg_hba.conf.sample 26 Aug 2004 16:50:05 -0000 1.53
--- src/backend/libpq/pg_hba.conf.sample 22 Sep 2004 14:17:41 -0000
***************
*** 60,67 ****
# TYPE DATABASE USER CIDR-ADDRESS METHOD
! local all all @authmethod@
! # IPv4-style local connections:
host all all 127.0.0.1/32 @authmethod@
! # IPv6-style local connections:
host all all ::1/128 @authmethod@
--- 60,68 ----
# TYPE DATABASE USER CIDR-ADDRESS METHOD
! @remove-line-for-win32@# "local" is for Unix domain socket connections only
! @remove-line-for-win32@local all all @authmethod@
! # IPv4 local connections:
host all all 127.0.0.1/32 @authmethod@
! # IPv6 local connections:
host all all ::1/128 @authmethod@
Index: src/bin/initdb/initdb.c
===================================================================
RCS file: /projects/cvsroot/pgsql-server/src/bin/initdb/initdb.c,v
retrieving revision 1.54
diff -c -w -r1.54 initdb.c
*** src/bin/initdb/initdb.c 2 Sep 2004 17:58:41 -0000 1.54
--- src/bin/initdb/initdb.c 22 Sep 2004 14:17:42 -0000
***************
*** 147,152 ****
--- 147,153 ----
static void *xmalloc(size_t size);
static char *xstrdup(const char *s);
static char **replace_token(char **lines, char *token, char *replacement);
+ static char **filter_lines_with_token(char **lines, char *token);
static char **readfile(char *path);
static void writefile(char *path, char **lines);
static int mkdir_p(char *path, mode_t omode);
***************
*** 311,316 ****
--- 312,348 ----
}
/*
+ * make a copy of lines without any that contain the token
+ * a sort of poor man's grep -v
+ *
+ */
+
+ static char **
+ filter_lines_with_token(char **lines, char *token)
+ {
+ int numlines = 1;
+ int i, src, dst;
+ char **result;
+
+ for (i = 0; lines[i]; i++)
+ numlines++;
+
+ result = (char **) xmalloc(numlines * sizeof(char *));
+
+ for (src = 0, dst = 0; src < numlines; src++)
+ {
+
+ if (lines[src] == NULL || strstr(lines[src], token) == NULL)
+ {
+ result[dst++] = lines[src];
+ }
+
+ }
+
+ return result;
+ }
+
+ /*
* get the lines from a text file
*/
static char **
***************
*** 1093,1098 ****
--- 1125,1136 ----
conflines = readfile(hba_file);
+ #ifdef WIN32
+ conflines = filter_lines_with_token(conflines,"@remove-line-for-win32@");
+ #else
+ conflines = replace_token(conflines,"@remove-line-for-win32@","");
+ #endif
+
#ifndef HAVE_IPV6
conflines = replace_token(conflines,
"host all all ::1",
Patch applied. Thanks. I adjusted the code slightly and I tested !HAVE_UNIX_SOCKETS rather than Win32. --------------------------------------------------------------------------- Andrew Dunstan wrote: > > The original of this seems to have gotten lost in the ether somewhere. > It might turn up some day ... > > andrew > > Tom Lane wrote: > > >Andrew Dunstan <andrew@dunslane.net> writes: > > > > > >>The attached tiny patch will possibly help to avoid some confusion by > >>Windows users about the "local" line in pg_hba.conf (and thus help > >>reduce queries to us ;-) ). > >> > >> > > > >I was wondering if we could teach initdb to remove that line altogether > >in Windows installations. > > > > > > > > > > I think this does what Tom and Peter suggested. I don't have a Windows > box to test it on, though. > > cheers > > andrew > > > > ---------------------------(end of broadcast)--------------------------- > TIP 7: don't forget to increase your free space map settings -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073