mirrors
/
postgres
mirror of https://github.com/postgres/postgres
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
postgres/contrib/btree_gist/btree_utils_var.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

613 lines
14 KiB
C
Raw Normal View History

Add $PostgreSQL$ markers to a lot of files that were missing them. This particular batch was just for *.c and *.h file. The changes were made with the following 2 commands: find . \( \( -name 'libstemmer' -o -name 'expected' -o -name 'ppport.h' \) -prune \) -o \( -name '*.[ch]' \) \( -exec grep -q '\$PostgreSQL' {} \; -o -print \) | while read file ; do head -n 1 < $file | grep -q '^/\*' && echo $file; done | xargs -l sed -i -e '1s/^\// /' -e '1i/*\n * $PostgreSQL:$ \n *' find . \( \( -name 'libstemmer' -o -name 'expected' -o -name 'ppport.h' \) -prune \) -o \( -name '*.[ch]' \) \( -exec grep -q '\$PostgreSQL' {} \; -o -print \) | xargs -l sed -i -e '1i/*\n * $PostgreSQL:$ \n */'
2008-05-17 05:28:26 +04:00
/*
Remove cvs keywords from all files.
2010-09-21 00:08:53 +04:00
* contrib/btree_gist/btree_utils_var.c
Add $PostgreSQL$ markers to a lot of files that were missing them. This particular batch was just for *.c and *.h file. The changes were made with the following 2 commands: find . \( \( -name 'libstemmer' -o -name 'expected' -o -name 'ppport.h' \) -prune \) -o \( -name '*.[ch]' \) \( -exec grep -q '\$PostgreSQL' {} \; -o -print \) | while read file ; do head -n 1 < $file | grep -q '^/\*' && echo $file; done | xargs -l sed -i -e '1s/^\// /' -e '1i/*\n * $PostgreSQL:$ \n *' find . \( \( -name 'libstemmer' -o -name 'expected' -o -name 'ppport.h' \) -prune \) -o \( -name '*.[ch]' \) \( -exec grep -q '\$PostgreSQL' {} \; -o -print \) | xargs -l sed -i -e '1i/*\n * $PostgreSQL:$ \n */'
2008-05-17 05:28:26 +04:00
*/
Add postgres.h to *.c files for pg_upgrade, ltree, and btree_gist, and remove from local *.h files. Per suggestion from Alvaro.
2011-08-27 05:16:24 +04:00
#include "postgres.h"
Adjust /contrib for new include file contents.
2006-07-13 20:57:31 +04:00
#include <math.h>
Add additional includes needed on some platforms.
2006-07-14 08:44:46 +04:00
#include <limits.h>
#include <float.h>
Adjust /contrib for new include file contents.
2006-07-13 20:57:31 +04:00
Make the order of the header file includes consistent in contrib modules. The basic rule we follow here is to always first include 'postgres.h' or 'postgres_fe.h' whichever is applicable, then system header includes and then Postgres header includes.  In this, we also follow that all the Postgres header includes are in order based on their ASCII value.  We generally follow these rules, but the code has deviated in many places. This commit makes it consistent just for contrib modules. The later commits will enforce similar rules in other parts of code. Author: Vignesh C Reviewed-by: Amit Kapila Discussion: https://postgr.es/m/CALDaNm2Sznv8RR6Ex-iJO6xAdsxgWhCoETkaYX=+9DW3q0QCfA@mail.gmail.com
2019-10-23 06:56:22 +03:00
#include "btree_gist.h"
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
#include "btree_utils_var.h"
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
#include "utils/builtins.h"
Make the order of the header file includes consistent in contrib modules. The basic rule we follow here is to always first include 'postgres.h' or 'postgres_fe.h' whichever is applicable, then system header includes and then Postgres header includes.  In this, we also follow that all the Postgres header includes are in order based on their ASCII value.  We generally follow these rules, but the code has deviated in many places. This commit makes it consistent just for contrib modules. The later commits will enforce similar rules in other parts of code. Author: Vignesh C Reviewed-by: Amit Kapila Discussion: https://postgr.es/m/CALDaNm2Sznv8RR6Ex-iJO6xAdsxgWhCoETkaYX=+9DW3q0QCfA@mail.gmail.com
2019-10-23 06:56:22 +03:00
#include "utils/pg_locale.h"
Improve our #include situation by moving pointer types away from the corresponding struct definitions. This allows other headers to avoid including certain highly-loaded headers such as rel.h and relscan.h, instead using just relcache.h, heapam.h or genam.h, which are more lightweight and thus cause less unnecessary dependencies.
2008-06-19 04:46:06 +04:00
#include "utils/rel.h"
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
/* used for key sorting */
typedef struct
{
int i;
GBT_VARKEY *t;
} Vsrt;
typedef struct
{
const gbtree_vinfo *tinfo;
Oid collation;
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
FmgrInfo *flinfo;
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
} gbt_vsrt_arg;
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
PG_FUNCTION_INFO_V1(gbt_var_decompress);
Add index-only scan support to btree_gist. inet, cidr, and timetz indexes still cannot support index-only scans, because they don't store the original unmodified value in the index, but a derived approximate value.
2015-03-28 00:35:16 +03:00
PG_FUNCTION_INFO_V1(gbt_var_fetch);
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
Datum
gbt_var_decompress(PG_FUNCTION_ARGS)
{
Cleanup of GiST extensions in contrib/: now that we always invoke GiST methods in a short-lived memory context, there is no need for GiST methods to do their own manual (and error-prone) memory management.
2005-05-21 16:08:06 +04:00
GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0);
Make DatumGetFoo/PG_GETARG_FOO/PG_RETURN_FOO macro names more consistent. By project convention, these names should include "P" when dealing with a pointer type; that is, if the result of a GETARG macro is of type FOO *, it should be called PG_GETARG_FOO_P not just PG_GETARG_FOO. Some newer types such as JSONB and ranges had not followed the convention, and a number of contrib modules hadn't gotten that memo either. Rename the offending macros to improve consistency. In passing, fix a few places that thought PG_DETOAST_DATUM() returns a Datum; it does not, it returns "struct varlena *". Applying DatumGetPointer to that happens not to cause any bad effects today, but it's formally wrong. Also, adjust an ltree macro that was designed without any thought for what pgindent would do with it. This is all cosmetic and shouldn't have any impact on generated code. Mark Dilger, some further tweaks by me Discussion: https://postgr.es/m/EA5676F4-766F-4F38-8348-ECC7DB427C6A@gmail.com
2017-09-18 22:21:23 +03:00
GBT_VARKEY *key = (GBT_VARKEY *) PG_DETOAST_DATUM(entry->key);
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
Cleanup of GiST extensions in contrib/: now that we always invoke GiST methods in a short-lived memory context, there is no need for GiST methods to do their own manual (and error-prone) memory management.
2005-05-21 16:08:06 +04:00
if (key != (GBT_VARKEY *) DatumGetPointer(entry->key))
{
GISTENTRY *retval = (GISTENTRY *) palloc(sizeof(GISTENTRY));
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
Cleanup of GiST extensions in contrib/: now that we always invoke GiST methods in a short-lived memory context, there is no need for GiST methods to do their own manual (and error-prone) memory management.
2005-05-21 16:08:06 +04:00
gistentryinit(*retval, PointerGetDatum(key),
entry->rel, entry->page,
Change TRUE/FALSE to true/false The lower case spellings are C and C++ standard and are used in most parts of the PostgreSQL sources. The upper case spellings are only used in some files/modules. So standardize on the standard spellings. The APIs for ICU, Perl, and Windows define their own TRUE and FALSE, so those are left as is when using those APIs. In code comments, we use the lower-case spelling for the C concepts and keep the upper-case spelling for the SQL concepts. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
2017-08-16 07:22:32 +03:00
entry->offset, false);
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
Cleanup of GiST extensions in contrib/: now that we always invoke GiST methods in a short-lived memory context, there is no need for GiST methods to do their own manual (and error-prone) memory management.
2005-05-21 16:08:06 +04:00
PG_RETURN_POINTER(retval);
}
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
Cleanup of GiST extensions in contrib/: now that we always invoke GiST methods in a short-lived memory context, there is no need for GiST methods to do their own manual (and error-prone) memory management.
2005-05-21 16:08:06 +04:00
PG_RETURN_POINTER(entry);
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
}
Fix misc typos. Oskari Saarenmaa. Backpatch to stable branches where applicable.
2015-09-05 11:35:49 +03:00
/* Returns a better readable representation of variable key ( sets pointer ) */
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
GBT_VARKEY_R
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
gbt_var_key_readable(const GBT_VARKEY *k)
{
GBT_VARKEY_R r;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
r.lower = (bytea *) &(((char *) k)[VARHDRSZ]);
if (VARSIZE(k) > (VARHDRSZ + (VARSIZE(r.lower))))
- Add aligment of variable data types - Add aligment for interval data types - Avoid floating point overflow in penalty functions Janko Richter <jankorichter@yahoo.de> and teodor
2004-06-03 16:26:10 +04:00
r.upper = (bytea *) &(((char *) k)[VARHDRSZ + INTALIGN(VARSIZE(r.lower))]);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
else
r.upper = r.lower;
return r;
}
Minor refactoring of btree_gist code. The gbt_var_key_copy function was doing two different things depending on the boolean argument. Seems cleaner to have two separate functions. Remove unused argument from gbt_num_compress.
2015-03-27 00:10:10 +03:00
/*
* Create a leaf-entry to store in the index, from a single Datum.
*/
static GBT_VARKEY *
gbt_var_key_from_datum(const struct varlena *u)
{
int32 lowersize = VARSIZE(u);
GBT_VARKEY *r;
r = (GBT_VARKEY *) palloc(lowersize + VARHDRSZ);
memcpy(VARDATA(r), u, lowersize);
SET_VARSIZE(r, lowersize + VARHDRSZ);
return r;
}
/*
* Create an entry to store in the index, from lower and upper bound.
*/
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
GBT_VARKEY *
Minor refactoring of btree_gist code. The gbt_var_key_copy function was doing two different things depending on the boolean argument. Seems cleaner to have two separate functions. Remove unused argument from gbt_num_compress.
2015-03-27 00:10:10 +03:00
gbt_var_key_copy(const GBT_VARKEY_R *u)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
Suppress some more valgrind whining about btree_gist. A couple of functions didn't bother to zero out pad bytes in datums that would ultimately go to disk. Harmless, but valgrind doesn't know that.
2014-05-16 23:29:02 +04:00
int32 lowersize = VARSIZE(u->lower);
int32 uppersize = VARSIZE(u->upper);
Minor refactoring of btree_gist code. The gbt_var_key_copy function was doing two different things depending on the boolean argument. Seems cleaner to have two separate functions. Remove unused argument from gbt_num_compress.
2015-03-27 00:10:10 +03:00
GBT_VARKEY *r;
r = (GBT_VARKEY *) palloc0(INTALIGN(lowersize) + uppersize + VARHDRSZ);
memcpy(VARDATA(r), u->lower, lowersize);
memcpy(VARDATA(r) + INTALIGN(lowersize), u->upper, uppersize);
SET_VARSIZE(r, INTALIGN(lowersize) + uppersize + VARHDRSZ);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
return r;
}
static GBT_VARKEY *
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
gbt_var_leaf2node(GBT_VARKEY *leaf, const gbtree_vinfo *tinfo, FmgrInfo *flinfo)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
GBT_VARKEY *out = leaf;
if (tinfo->f_l2n)
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
out = tinfo->f_l2n(leaf, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
return out;
}
/*
* returns the common prefix length of a node key
*/
static int32
gbt_var_node_cp_len(const GBT_VARKEY *node, const gbtree_vinfo *tinfo)
{
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
GBT_VARKEY_R r = gbt_var_key_readable(node);
int32 i = 0;
int32 l = 0;
int32 t1len = VARSIZE(r.lower) - VARHDRSZ;
int32 t2len = VARSIZE(r.upper) - VARHDRSZ;
int32 ml = Min(t1len, t2len);
char *p1 = VARDATA(r.lower);
char *p2 = VARDATA(r.upper);
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
if (ml == 0)
return 0;
while (i < ml)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
if (tinfo->eml > 1 && l == 0)
{
if ((l = pg_mblen(p1)) != pg_mblen(p2))
{
return i;
Standard pgindent run for 8.1.
2005-10-15 06:49:52 +04:00
}
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
}
if (*p1 != *p2)
{
if (tinfo->eml > 1)
{
return (i - l + 1);
}
else
{
return i;
}
Standard pgindent run for 8.1.
2005-10-15 06:49:52 +04:00
}
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
p1++;
p2++;
l--;
i++;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
Make DatumGetFoo/PG_GETARG_FOO/PG_RETURN_FOO macro names more consistent. By project convention, these names should include "P" when dealing with a pointer type; that is, if the result of a GETARG macro is of type FOO *, it should be called PG_GETARG_FOO_P not just PG_GETARG_FOO. Some newer types such as JSONB and ranges had not followed the convention, and a number of contrib modules hadn't gotten that memo either. Rename the offending macros to improve consistency. In passing, fix a few places that thought PG_DETOAST_DATUM() returns a Datum; it does not, it returns "struct varlena *". Applying DatumGetPointer to that happens not to cause any bad effects today, but it's formally wrong. Also, adjust an ltree macro that was designed without any thought for what pgindent would do with it. This is all cosmetic and shouldn't have any impact on generated code. Mark Dilger, some further tweaks by me Discussion: https://postgr.es/m/EA5676F4-766F-4F38-8348-ECC7DB427C6A@gmail.com
2017-09-18 22:21:23 +03:00
return ml; /* lower == upper */
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
/*
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
* returns true, if query matches prefix ( common prefix )
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
*/
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
static bool
gbt_bytea_pf_match(const bytea *pf, const bytea *query, const gbtree_vinfo *tinfo)
{
Change TRUE/FALSE to true/false The lower case spellings are C and C++ standard and are used in most parts of the PostgreSQL sources. The upper case spellings are only used in some files/modules. So standardize on the standard spellings. The APIs for ICU, Perl, and Windows define their own TRUE and FALSE, so those are left as is when using those APIs. In code comments, we use the lower-case spelling for the C concepts and keep the upper-case spelling for the SQL concepts. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
2017-08-16 07:22:32 +03:00
bool out = false;
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
int32 qlen = VARSIZE(query) - VARHDRSZ;
int32 nlen = VARSIZE(pf) - VARHDRSZ;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
if (nlen <= qlen)
{
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
char *q = VARDATA(query);
char *n = VARDATA(pf);
Standard pgindent run for 8.1.
2005-10-15 06:49:52 +04:00
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
out = (memcmp(q, n, nlen) == 0);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
return out;
}
/*
* returns true, if query matches node using common prefix
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
*/
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
static bool
gbt_var_node_pf_match(const GBT_VARKEY_R *node, const bytea *query, const gbtree_vinfo *tinfo)
{
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
return (tinfo->trnc &&
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
(gbt_bytea_pf_match(node->lower, query, tinfo) ||
gbt_bytea_pf_match(node->upper, query, tinfo)));
}
/*
* truncates / compresses the node key
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
* cpf_length .. common prefix length
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
*/
static GBT_VARKEY *
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
gbt_var_node_truncate(const GBT_VARKEY *node, int32 cpf_length, const gbtree_vinfo *tinfo)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
GBT_VARKEY *out = NULL;
GBT_VARKEY_R r = gbt_var_key_readable(node);
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
int32 len1 = VARSIZE(r.lower) - VARHDRSZ;
int32 len2 = VARSIZE(r.upper) - VARHDRSZ;
Replace direct assignments to VARATT_SIZEP(x) with SET_VARSIZE(x, len). Get rid of VARATT_SIZE and VARATT_DATA, which were simply redundant with VARSIZE and VARDATA, and as a consequence almost no code was using the longer names. Rename the length fields of struct varlena and various derived structures to catch anyplace that was accessing them directly; and clean up various places so caught. In itself this patch doesn't change any behavior at all, but it is necessary infrastructure if we hope to play any games with the representation of varlena headers. Greg Stark and Tom Lane
2007-02-28 02:48:10 +03:00
int32 si;
char *out2;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
len1 = Min(len1, (cpf_length + 1));
len2 = Min(len2, (cpf_length + 1));
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Replace direct assignments to VARATT_SIZEP(x) with SET_VARSIZE(x, len). Get rid of VARATT_SIZE and VARATT_DATA, which were simply redundant with VARSIZE and VARDATA, and as a consequence almost no code was using the longer names. Rename the length fields of struct varlena and various derived structures to catch anyplace that was accessing them directly; and clean up various places so caught. In itself this patch doesn't change any behavior at all, but it is necessary infrastructure if we hope to play any games with the representation of varlena headers. Greg Stark and Tom Lane
2007-02-28 02:48:10 +03:00
si = 2 * VARHDRSZ + INTALIGN(len1 + VARHDRSZ) + len2;
Suppress some more valgrind whining about btree_gist. A couple of functions didn't bother to zero out pad bytes in datums that would ultimately go to disk. Harmless, but valgrind doesn't know that.
2014-05-16 23:29:02 +04:00
out = (GBT_VARKEY *) palloc0(si);
Replace direct assignments to VARATT_SIZEP(x) with SET_VARSIZE(x, len). Get rid of VARATT_SIZE and VARATT_DATA, which were simply redundant with VARSIZE and VARDATA, and as a consequence almost no code was using the longer names. Rename the length fields of struct varlena and various derived structures to catch anyplace that was accessing them directly; and clean up various places so caught. In itself this patch doesn't change any behavior at all, but it is necessary infrastructure if we hope to play any games with the representation of varlena headers. Greg Stark and Tom Lane
2007-02-28 02:48:10 +03:00
SET_VARSIZE(out, si);
memcpy(VARDATA(out), r.lower, len1 + VARHDRSZ);
SET_VARSIZE(VARDATA(out), len1 + VARHDRSZ);
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Replace direct assignments to VARATT_SIZEP(x) with SET_VARSIZE(x, len). Get rid of VARATT_SIZE and VARATT_DATA, which were simply redundant with VARSIZE and VARDATA, and as a consequence almost no code was using the longer names. Rename the length fields of struct varlena and various derived structures to catch anyplace that was accessing them directly; and clean up various places so caught. In itself this patch doesn't change any behavior at all, but it is necessary infrastructure if we hope to play any games with the representation of varlena headers. Greg Stark and Tom Lane
2007-02-28 02:48:10 +03:00
out2 = VARDATA(out) + INTALIGN(len1 + VARHDRSZ);
memcpy(out2, r.upper, len2 + VARHDRSZ);
SET_VARSIZE(out2, len2 + VARHDRSZ);
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
return out;
}
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
void
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_var_bin_union(Datum *u, GBT_VARKEY *e, Oid collation,
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
const gbtree_vinfo *tinfo, FmgrInfo *flinfo)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
GBT_VARKEY_R eo = gbt_var_key_readable(e);
Fix erroneous range-union logic for varlena types in contrib/btree_gist. gbt_var_bin_union() failed to do the right thing when the existing range needed to be widened at both ends rather than just one end. This could result in an invalid index in which keys that are present would not be found by searches, because the searches would not think they need to descend to the relevant leaf pages. This error affected all the varlena datatypes supported by btree_gist (text, bytea, bit, numeric). Per investigation of a trouble report from Tomas Vondra. (There is also an issue in gbt_var_penalty(), but that should only result in inefficiency not wrong answers. I'm committing this separately so that we have a git state in which it can be tested that bad penalty results don't produce invalid indexes.) Back-patch to all supported branches.
2013-02-08 03:22:19 +04:00
GBT_VARKEY_R nr;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
if (eo.lower == eo.upper) /* leaf */
{
Fix erroneous range-union logic for varlena types in contrib/btree_gist. gbt_var_bin_union() failed to do the right thing when the existing range needed to be widened at both ends rather than just one end. This could result in an invalid index in which keys that are present would not be found by searches, because the searches would not think they need to descend to the relevant leaf pages. This error affected all the varlena datatypes supported by btree_gist (text, bytea, bit, numeric). Per investigation of a trouble report from Tomas Vondra. (There is also an issue in gbt_var_penalty(), but that should only result in inefficiency not wrong answers. I'm committing this separately so that we have a git state in which it can be tested that bad penalty results don't produce invalid indexes.) Back-patch to all supported branches.
2013-02-08 03:22:19 +04:00
GBT_VARKEY *tmp;
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
tmp = gbt_var_leaf2node(e, tinfo, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
if (tmp != e)
eo = gbt_var_key_readable(tmp);
}
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
if (DatumGetPointer(*u))
{
GBT_VARKEY_R ro = gbt_var_key_readable((GBT_VARKEY *) DatumGetPointer(*u));
Fix erroneous range-union logic for varlena types in contrib/btree_gist. gbt_var_bin_union() failed to do the right thing when the existing range needed to be widened at both ends rather than just one end. This could result in an invalid index in which keys that are present would not be found by searches, because the searches would not think they need to descend to the relevant leaf pages. This error affected all the varlena datatypes supported by btree_gist (text, bytea, bit, numeric). Per investigation of a trouble report from Tomas Vondra. (There is also an issue in gbt_var_penalty(), but that should only result in inefficiency not wrong answers. I'm committing this separately so that we have a git state in which it can be tested that bad penalty results don't produce invalid indexes.) Back-patch to all supported branches.
2013-02-08 03:22:19 +04:00
bool update = false;
nr.lower = ro.lower;
nr.upper = ro.upper;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
if (tinfo->f_cmp(ro.lower, eo.lower, collation, flinfo) > 0)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
nr.lower = eo.lower;
Fix erroneous range-union logic for varlena types in contrib/btree_gist. gbt_var_bin_union() failed to do the right thing when the existing range needed to be widened at both ends rather than just one end. This could result in an invalid index in which keys that are present would not be found by searches, because the searches would not think they need to descend to the relevant leaf pages. This error affected all the varlena datatypes supported by btree_gist (text, bytea, bit, numeric). Per investigation of a trouble report from Tomas Vondra. (There is also an issue in gbt_var_penalty(), but that should only result in inefficiency not wrong answers. I'm committing this separately so that we have a git state in which it can be tested that bad penalty results don't produce invalid indexes.) Back-patch to all supported branches.
2013-02-08 03:22:19 +04:00
update = true;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
Cleanup of GiST extensions in contrib/: now that we always invoke GiST methods in a short-lived memory context, there is no need for GiST methods to do their own manual (and error-prone) memory management.
2005-05-21 16:08:06 +04:00
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
if (tinfo->f_cmp(ro.upper, eo.upper, collation, flinfo) < 0)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
nr.upper = eo.upper;
Fix erroneous range-union logic for varlena types in contrib/btree_gist. gbt_var_bin_union() failed to do the right thing when the existing range needed to be widened at both ends rather than just one end. This could result in an invalid index in which keys that are present would not be found by searches, because the searches would not think they need to descend to the relevant leaf pages. This error affected all the varlena datatypes supported by btree_gist (text, bytea, bit, numeric). Per investigation of a trouble report from Tomas Vondra. (There is also an issue in gbt_var_penalty(), but that should only result in inefficiency not wrong answers. I'm committing this separately so that we have a git state in which it can be tested that bad penalty results don't produce invalid indexes.) Back-patch to all supported branches.
2013-02-08 03:22:19 +04:00
update = true;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
Cleanup of GiST extensions in contrib/: now that we always invoke GiST methods in a short-lived memory context, there is no need for GiST methods to do their own manual (and error-prone) memory management.
2005-05-21 16:08:06 +04:00
Fix erroneous range-union logic for varlena types in contrib/btree_gist. gbt_var_bin_union() failed to do the right thing when the existing range needed to be widened at both ends rather than just one end. This could result in an invalid index in which keys that are present would not be found by searches, because the searches would not think they need to descend to the relevant leaf pages. This error affected all the varlena datatypes supported by btree_gist (text, bytea, bit, numeric). Per investigation of a trouble report from Tomas Vondra. (There is also an issue in gbt_var_penalty(), but that should only result in inefficiency not wrong answers. I'm committing this separately so that we have a git state in which it can be tested that bad penalty results don't produce invalid indexes.) Back-patch to all supported branches.
2013-02-08 03:22:19 +04:00
if (update)
Minor refactoring of btree_gist code. The gbt_var_key_copy function was doing two different things depending on the boolean argument. Seems cleaner to have two separate functions. Remove unused argument from gbt_num_compress.
2015-03-27 00:10:10 +03:00
*u = PointerGetDatum(gbt_var_key_copy(&nr));
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
else
{
nr.lower = eo.lower;
nr.upper = eo.upper;
Minor refactoring of btree_gist code. The gbt_var_key_copy function was doing two different things depending on the boolean argument. Seems cleaner to have two separate functions. Remove unused argument from gbt_num_compress.
2015-03-27 00:10:10 +03:00
*u = PointerGetDatum(gbt_var_key_copy(&nr));
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
}
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
GISTENTRY *
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
gbt_var_compress(GISTENTRY *entry, const gbtree_vinfo *tinfo)
{
GISTENTRY *retval;
if (entry->leafkey)
{
Minor refactoring of btree_gist code. The gbt_var_key_copy function was doing two different things depending on the boolean argument. Seems cleaner to have two separate functions. Remove unused argument from gbt_num_compress.
2015-03-27 00:10:10 +03:00
struct varlena *leaf = PG_DETOAST_DATUM(entry->key);
GBT_VARKEY *r;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
Minor refactoring of btree_gist code. The gbt_var_key_copy function was doing two different things depending on the boolean argument. Seems cleaner to have two separate functions. Remove unused argument from gbt_num_compress.
2015-03-27 00:10:10 +03:00
r = gbt_var_key_from_datum(leaf);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
retval = palloc(sizeof(GISTENTRY));
gistentryinit(*retval, PointerGetDatum(r),
entry->rel, entry->page,
Change TRUE/FALSE to true/false The lower case spellings are C and C++ standard and are used in most parts of the PostgreSQL sources. The upper case spellings are only used in some files/modules. So standardize on the standard spellings. The APIs for ICU, Perl, and Windows define their own TRUE and FALSE, so those are left as is when using those APIs. In code comments, we use the lower-case spelling for the C concepts and keep the upper-case spelling for the SQL concepts. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
2017-08-16 07:22:32 +03:00
entry->offset, true);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
else
retval = entry;
Remove unnecessary parentheses in return statements The parenthesized style has only been used in a few modules. Change that to use the style that is predominant across the whole tree. Reviewed-by: Michael Paquier <michael.paquier@gmail.com> Reviewed-by: Ryan Murphy <ryanfmurphy@gmail.com>
2017-08-17 19:39:20 +03:00
return retval;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
Add index-only scan support to btree_gist. inet, cidr, and timetz indexes still cannot support index-only scans, because they don't store the original unmodified value in the index, but a derived approximate value.
2015-03-28 00:35:16 +03:00
Datum
gbt_var_fetch(PG_FUNCTION_ARGS)
{
GISTENTRY *entry = (GISTENTRY *) PG_GETARG_POINTER(0);
Make DatumGetFoo/PG_GETARG_FOO/PG_RETURN_FOO macro names more consistent. By project convention, these names should include "P" when dealing with a pointer type; that is, if the result of a GETARG macro is of type FOO *, it should be called PG_GETARG_FOO_P not just PG_GETARG_FOO. Some newer types such as JSONB and ranges had not followed the convention, and a number of contrib modules hadn't gotten that memo either. Rename the offending macros to improve consistency. In passing, fix a few places that thought PG_DETOAST_DATUM() returns a Datum; it does not, it returns "struct varlena *". Applying DatumGetPointer to that happens not to cause any bad effects today, but it's formally wrong. Also, adjust an ltree macro that was designed without any thought for what pgindent would do with it. This is all cosmetic and shouldn't have any impact on generated code. Mark Dilger, some further tweaks by me Discussion: https://postgr.es/m/EA5676F4-766F-4F38-8348-ECC7DB427C6A@gmail.com
2017-09-18 22:21:23 +03:00
GBT_VARKEY *key = (GBT_VARKEY *) PG_DETOAST_DATUM(entry->key);
Add index-only scan support to btree_gist. inet, cidr, and timetz indexes still cannot support index-only scans, because they don't store the original unmodified value in the index, but a derived approximate value.
2015-03-28 00:35:16 +03:00
GBT_VARKEY_R r = gbt_var_key_readable(key);
GISTENTRY *retval;
retval = palloc(sizeof(GISTENTRY));
gistentryinit(*retval, PointerGetDatum(r.lower),
entry->rel, entry->page,
Change TRUE/FALSE to true/false The lower case spellings are C and C++ standard and are used in most parts of the PostgreSQL sources. The upper case spellings are only used in some files/modules. So standardize on the standard spellings. The APIs for ICU, Perl, and Windows define their own TRUE and FALSE, so those are left as is when using those APIs. In code comments, we use the lower-case spelling for the C concepts and keep the upper-case spelling for the SQL concepts. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
2017-08-16 07:22:32 +03:00
entry->offset, true);
Add index-only scan support to btree_gist. inet, cidr, and timetz indexes still cannot support index-only scans, because they don't store the original unmodified value in the index, but a derived approximate value.
2015-03-28 00:35:16 +03:00
PG_RETURN_POINTER(retval);
}
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
GBT_VARKEY *
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_var_union(const GistEntryVector *entryvec, int32 *size, Oid collation,
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
const gbtree_vinfo *tinfo, FmgrInfo *flinfo)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
int i = 0,
numranges = entryvec->n;
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
GBT_VARKEY *cur;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
Datum out;
GBT_VARKEY_R rk;
*size = sizeof(GBT_VARKEY);
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
cur = (GBT_VARKEY *) DatumGetPointer(entryvec->vector[0].key);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
rk = gbt_var_key_readable(cur);
Minor refactoring of btree_gist code. The gbt_var_key_copy function was doing two different things depending on the boolean argument. Seems cleaner to have two separate functions. Remove unused argument from gbt_num_compress.
2015-03-27 00:10:10 +03:00
out = PointerGetDatum(gbt_var_key_copy(&rk));
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
for (i = 1; i < numranges; i++)
{
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
cur = (GBT_VARKEY *) DatumGetPointer(entryvec->vector[i].key);
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
gbt_var_bin_union(&out, cur, collation, tinfo, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
/* Truncate (=compress) key */
if (tinfo->trnc)
{
int32 plen;
GBT_VARKEY *trc = NULL;
plen = gbt_var_node_cp_len((GBT_VARKEY *) DatumGetPointer(out), tinfo);
trc = gbt_var_node_truncate((GBT_VARKEY *) DatumGetPointer(out), plen + 1, tinfo);
out = PointerGetDatum(trc);
}
return ((GBT_VARKEY *) DatumGetPointer(out));
}
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
bool
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_var_same(Datum d1, Datum d2, Oid collation,
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
const gbtree_vinfo *tinfo, FmgrInfo *flinfo)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
GBT_VARKEY *t1 = (GBT_VARKEY *) DatumGetPointer(d1);
GBT_VARKEY *t2 = (GBT_VARKEY *) DatumGetPointer(d2);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
GBT_VARKEY_R r1,
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
r2;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
r1 = gbt_var_key_readable(t1);
r2 = gbt_var_key_readable(t2);
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
return (tinfo->f_cmp(r1.lower, r2.lower, collation, flinfo) == 0 &&
tinfo->f_cmp(r1.upper, r2.upper, collation, flinfo) == 0);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
float *
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_var_penalty(float *res, const GISTENTRY *o, const GISTENTRY *n,
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
Oid collation, const gbtree_vinfo *tinfo, FmgrInfo *flinfo)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
GBT_VARKEY *orge = (GBT_VARKEY *) DatumGetPointer(o->key);
GBT_VARKEY *newe = (GBT_VARKEY *) DatumGetPointer(n->key);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
GBT_VARKEY_R ok,
nk;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
*res = 0.0;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
nk = gbt_var_key_readable(newe);
if (nk.lower == nk.upper) /* leaf */
{
Make contrib/btree_gist's GiST penalty function a bit saner. The previous coding supposed that the first differing bytes in two varlena datums must have the same sign difference as their overall comparison result. This is obviously bogus for text strings in non-C locales, and probably wrong for numeric, and even for bytea I think it was wrong on machines where char is signed. When the assumption failed, the function could deliver a zero or negative penalty in situations where such a result is quite ridiculous, leading the core GiST code to make very bad page-split decisions. To fix, take the absolute values of the byte-level differences. Also, switch the code to using unsigned char not just char, so that the behavior will be consistent whether char is signed or not. Per investigation of a trouble report from Tomas Vondra. Back-patch to all supported branches.
2013-02-08 04:13:59 +04:00
GBT_VARKEY *tmp;
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
tmp = gbt_var_leaf2node(newe, tinfo, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
if (tmp != newe)
nk = gbt_var_key_readable(tmp);
}
ok = gbt_var_key_readable(orge);
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Fixes from Janko Richter <jankorichter@yahoo.de> - Fix wrong index results on text, char, varchar for multibyte strings - Fix some SIGFPE signals - Add support for infinite timestamps - Because of locale settings, btree_gist can not be a prefix index anymore (for text). Each node holds now just the lower and upper boundary.
2005-07-01 17:44:56 +04:00
if ((VARSIZE(ok.lower) - VARHDRSZ) == 0 && (VARSIZE(ok.upper) - VARHDRSZ) == 0)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
*res = 0.0;
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
else if (!((tinfo->f_cmp(nk.lower, ok.lower, collation, flinfo) >= 0 ||
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_bytea_pf_match(ok.lower, nk.lower, tinfo)) &&
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
(tinfo->f_cmp(nk.upper, ok.upper, collation, flinfo) <= 0 ||
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_bytea_pf_match(ok.upper, nk.upper, tinfo))))
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
Datum d = PointerGetDatum(0);
Make contrib/btree_gist's GiST penalty function a bit saner. The previous coding supposed that the first differing bytes in two varlena datums must have the same sign difference as their overall comparison result. This is obviously bogus for text strings in non-C locales, and probably wrong for numeric, and even for bytea I think it was wrong on machines where char is signed. When the assumption failed, the function could deliver a zero or negative penalty in situations where such a result is quite ridiculous, leading the core GiST code to make very bad page-split decisions. To fix, take the absolute values of the byte-level differences. Also, switch the code to using unsigned char not just char, so that the behavior will be consistent whether char is signed or not. Per investigation of a trouble report from Tomas Vondra. Back-patch to all supported branches.
2013-02-08 04:13:59 +04:00
double dres;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
int32 ol,
ul;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
gbt_var_bin_union(&d, orge, collation, tinfo, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
ol = gbt_var_node_cp_len((GBT_VARKEY *) DatumGetPointer(d), tinfo);
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
gbt_var_bin_union(&d, newe, collation, tinfo, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
ul = gbt_var_node_cp_len((GBT_VARKEY *) DatumGetPointer(d), tinfo);
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
if (ul < ol)
{
Make contrib/btree_gist's GiST penalty function a bit saner. The previous coding supposed that the first differing bytes in two varlena datums must have the same sign difference as their overall comparison result. This is obviously bogus for text strings in non-C locales, and probably wrong for numeric, and even for bytea I think it was wrong on machines where char is signed. When the assumption failed, the function could deliver a zero or negative penalty in situations where such a result is quite ridiculous, leading the core GiST code to make very bad page-split decisions. To fix, take the absolute values of the byte-level differences. Also, switch the code to using unsigned char not just char, so that the behavior will be consistent whether char is signed or not. Per investigation of a trouble report from Tomas Vondra. Back-patch to all supported branches.
2013-02-08 04:13:59 +04:00
dres = (ol - ul); /* reduction of common prefix len */
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
else
{
GBT_VARKEY_R uk = gbt_var_key_readable((GBT_VARKEY *) DatumGetPointer(d));
Make contrib/btree_gist's GiST penalty function a bit saner. The previous coding supposed that the first differing bytes in two varlena datums must have the same sign difference as their overall comparison result. This is obviously bogus for text strings in non-C locales, and probably wrong for numeric, and even for bytea I think it was wrong on machines where char is signed. When the assumption failed, the function could deliver a zero or negative penalty in situations where such a result is quite ridiculous, leading the core GiST code to make very bad page-split decisions. To fix, take the absolute values of the byte-level differences. Also, switch the code to using unsigned char not just char, so that the behavior will be consistent whether char is signed or not. Per investigation of a trouble report from Tomas Vondra. Back-patch to all supported branches.
2013-02-08 04:13:59 +04:00
unsigned char tmp[4];
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Make contrib/btree_gist's GiST penalty function a bit saner. The previous coding supposed that the first differing bytes in two varlena datums must have the same sign difference as their overall comparison result. This is obviously bogus for text strings in non-C locales, and probably wrong for numeric, and even for bytea I think it was wrong on machines where char is signed. When the assumption failed, the function could deliver a zero or negative penalty in situations where such a result is quite ridiculous, leading the core GiST code to make very bad page-split decisions. To fix, take the absolute values of the byte-level differences. Also, switch the code to using unsigned char not just char, so that the behavior will be consistent whether char is signed or not. Per investigation of a trouble report from Tomas Vondra. Back-patch to all supported branches.
2013-02-08 04:13:59 +04:00
tmp[0] = (unsigned char) (((VARSIZE(ok.lower) - VARHDRSZ) <= ul) ? 0 : (VARDATA(ok.lower)[ul]));
tmp[1] = (unsigned char) (((VARSIZE(uk.lower) - VARHDRSZ) <= ul) ? 0 : (VARDATA(uk.lower)[ul]));
tmp[2] = (unsigned char) (((VARSIZE(ok.upper) - VARHDRSZ) <= ul) ? 0 : (VARDATA(ok.upper)[ul]));
tmp[3] = (unsigned char) (((VARSIZE(uk.upper) - VARHDRSZ) <= ul) ? 0 : (VARDATA(uk.upper)[ul]));
Remove unnecessary uses of Abs() Use C standard abs() or fabs() instead. Reviewed-by: Zhang Mingli <zmlpostgres@gmail.com> Reviewed-by: Tom Lane <tgl@sss.pgh.pa.us> Discussion: https://www.postgresql.org/message-id/flat/4beb42b5-216b-bce8-d452-d924d5794c63%40enterprisedb.com
2022-10-07 14:28:38 +03:00
dres = abs(tmp[0] - tmp[1]) + abs(tmp[3] - tmp[2]);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
dres /= 256.0;
}
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
*res += FLT_MIN;
*res += (float) (dres / ((double) (ol + 1)));
*res *= (FLT_MAX / (o->rel->rd_att->natts + 1));
}
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
return res;
}
Whack btree_gist code around until it has some small hope of building on non-gcc compilers.
2004-08-21 04:09:10 +04:00
static int
Make use of qsort_arg in several places that were formerly using klugy static variables. This avoids any risk of potential non-reentrancy, and in particular offers a much cleaner workaround for the Intel compiler bug that was affecting ginutil.c.
2006-10-05 21:57:40 +04:00
gbt_vsrt_cmp(const void *a, const void *b, void *arg)
Whack btree_gist code around until it has some small hope of building on non-gcc compilers.
2004-08-21 04:09:10 +04:00
{
GBT_VARKEY_R ar = gbt_var_key_readable(((const Vsrt *) a)->t);
GBT_VARKEY_R br = gbt_var_key_readable(((const Vsrt *) b)->t);
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
const gbt_vsrt_arg *varg = (const gbt_vsrt_arg *) arg;
Preventing intersection of ranges during page split. Changes are only optimization, so don't backpatch.
2009-12-02 16:13:24 +03:00
int res;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
res = varg->tinfo->f_cmp(ar.lower, br.lower, varg->collation, varg->flinfo);
Preventing intersection of ranges during page split. Changes are only optimization, so don't backpatch.
2009-12-02 16:13:24 +03:00
if (res == 0)
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
return varg->tinfo->f_cmp(ar.upper, br.upper, varg->collation, varg->flinfo);
Preventing intersection of ranges during page split. Changes are only optimization, so don't backpatch.
2009-12-02 16:13:24 +03:00
return res;
Whack btree_gist code around until it has some small hope of building on non-gcc compilers.
2004-08-21 04:09:10 +04:00
}
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
GIST_SPLITVEC *
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_var_picksplit(const GistEntryVector *entryvec, GIST_SPLITVEC *v,
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
Oid collation, const gbtree_vinfo *tinfo, FmgrInfo *flinfo)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
Whack btree_gist code around until it has some small hope of building on non-gcc compilers.
2004-08-21 04:09:10 +04:00
OffsetNumber i,
maxoff = entryvec->n - 1;
Vsrt *arr;
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
int svcntr = 0,
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
nbytes;
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
char *cur;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
GBT_VARKEY **sv = NULL;
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_vsrt_arg varg;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Whack btree_gist code around until it has some small hope of building on non-gcc compilers.
2004-08-21 04:09:10 +04:00
arr = (Vsrt *) palloc((maxoff + 1) * sizeof(Vsrt));
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
nbytes = (maxoff + 2) * sizeof(OffsetNumber);
v->spl_left = (OffsetNumber *) palloc(nbytes);
v->spl_right = (OffsetNumber *) palloc(nbytes);
v->spl_ldatum = PointerGetDatum(0);
v->spl_rdatum = PointerGetDatum(0);
v->spl_nleft = 0;
v->spl_nright = 0;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
sv = palloc(sizeof(bytea *) * (maxoff + 1));
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
/* Sort entries */
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
for (i = FirstOffsetNumber; i <= maxoff; i = OffsetNumberNext(i))
{
GBT_VARKEY_R ro;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Simplify/clean up code for varlena types
2005-03-01 18:40:04 +03:00
cur = (char *) DatumGetPointer(entryvec->vector[i].key);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
ro = gbt_var_key_readable((GBT_VARKEY *) cur);
if (ro.lower == ro.upper) /* leaf */
{
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
sv[svcntr] = gbt_var_leaf2node((GBT_VARKEY *) cur, tinfo, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
arr[i].t = sv[svcntr];
if (sv[svcntr] != (GBT_VARKEY *) cur)
svcntr++;
}
else
arr[i].t = (GBT_VARKEY *) cur;
arr[i].i = i;
}
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
/* sort */
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
varg.tinfo = tinfo;
varg.collation = collation;
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
varg.flinfo = flinfo;
Make use of qsort_arg in several places that were formerly using klugy static variables. This avoids any risk of potential non-reentrancy, and in particular offers a much cleaner workaround for the Intel compiler bug that was affecting ginutil.c.
2006-10-05 21:57:40 +04:00
qsort_arg((void *) &arr[FirstOffsetNumber],
maxoff - FirstOffsetNumber + 1,
sizeof(Vsrt),
gbt_vsrt_cmp,
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
(void *) &varg);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
/* We do simply create two parts */
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
for (i = FirstOffsetNumber; i <= maxoff; i = OffsetNumberNext(i))
{
if (i <= (maxoff - FirstOffsetNumber + 1) / 2)
{
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
gbt_var_bin_union(&v->spl_ldatum, arr[i].t, collation, tinfo, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
v->spl_left[v->spl_nleft] = arr[i].i;
v->spl_nleft++;
}
else
{
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
gbt_var_bin_union(&v->spl_rdatum, arr[i].t, collation, tinfo, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
v->spl_right[v->spl_nright] = arr[i].i;
v->spl_nright++;
}
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
}
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
/* Truncate (=compress) key */
if (tinfo->trnc)
{
int32 ll = gbt_var_node_cp_len((GBT_VARKEY *) DatumGetPointer(v->spl_ldatum), tinfo);
int32 lr = gbt_var_node_cp_len((GBT_VARKEY *) DatumGetPointer(v->spl_rdatum), tinfo);
GBT_VARKEY *dl;
GBT_VARKEY *dr;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
ll = Max(ll, lr);
ll++;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
dl = gbt_var_node_truncate((GBT_VARKEY *) DatumGetPointer(v->spl_ldatum), ll, tinfo);
dr = gbt_var_node_truncate((GBT_VARKEY *) DatumGetPointer(v->spl_rdatum), ll, tinfo);
v->spl_ldatum = PointerGetDatum(dl);
v->spl_rdatum = PointerGetDatum(dr);
}
Whack btree_gist code around until it has some small hope of building on non-gcc compilers.
2004-08-21 04:09:10 +04:00
return v;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
/*
Whack btree_gist code around until it has some small hope of building on non-gcc compilers.
2004-08-21 04:09:10 +04:00
* The GiST consistent method
*/
This patch makes some minor style cleanups to contrib/btree_gist: remove the "extern" keyword from function definitions, reorganize some PG_GETARG_XXX() usage, and similar.
2005-05-12 04:39:37 +04:00
bool
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_var_consistent(GBT_VARKEY_R *key,
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
const void *query,
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
StrategyNumber strategy,
Oid collation,
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
bool is_leaf,
Use CallerFInfoFunctionCall with btree_gist for varlena types Follow up to commit 393bb504d7 which did this for numeric types.
2017-03-21 16:53:35 +03:00
const gbtree_vinfo *tinfo,
FmgrInfo *flinfo)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
Change TRUE/FALSE to true/false The lower case spellings are C and C++ standard and are used in most parts of the PostgreSQL sources. The upper case spellings are only used in some files/modules. So standardize on the standard spellings. The APIs for ICU, Perl, and Windows define their own TRUE and FALSE, so those are left as is when using those APIs. In code comments, we use the lower-case spelling for the C concepts and keep the upper-case spelling for the SQL concepts. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
2017-08-16 07:22:32 +03:00
bool retval = false;
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
switch (strategy)
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
{
case BTLessEqualStrategyNumber:
if (is_leaf)
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_ge(query, key->lower, collation, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
else
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_cmp(query, key->lower, collation, flinfo) >= 0
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
|| gbt_var_node_pf_match(key, query, tinfo);
break;
case BTLessStrategyNumber:
if (is_leaf)
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_gt(query, key->lower, collation, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
else
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_cmp(query, key->lower, collation, flinfo) >= 0
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
|| gbt_var_node_pf_match(key, query, tinfo);
break;
case BTEqualStrategyNumber:
if (is_leaf)
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_eq(query, key->lower, collation, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
else
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
retval =
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
(tinfo->f_cmp(key->lower, query, collation, flinfo) <= 0 &&
tinfo->f_cmp(query, key->upper, collation, flinfo) <= 0) ||
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
gbt_var_node_pf_match(key, query, tinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
break;
case BTGreaterStrategyNumber:
if (is_leaf)
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_lt(query, key->upper, collation, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
else
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_cmp(query, key->upper, collation, flinfo) <= 0
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
|| gbt_var_node_pf_match(key, query, tinfo);
break;
case BTGreaterEqualStrategyNumber:
if (is_leaf)
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_le(query, key->upper, collation, flinfo);
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
else
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = tinfo->f_cmp(query, key->upper, collation, flinfo) <= 0
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
|| gbt_var_node_pf_match(key, query, tinfo);
break;
Add btree_gist support for searching on "not equals". Jeff Davis, with slight editorialization by me.
2010-08-02 20:26:48 +04:00
case BtreeGistNotEqualStrategyNumber:
Reduce excessive dereferencing of function pointers It is equivalent in ANSI C to write (*funcptr) () and funcptr(). These two styles have been applied inconsistently. After discussion, we'll use the more verbose style for plain function pointer variables, to make it clear that it's a variable, and the shorter style when the function pointer is in a struct (s.func() or s->func()), because then it's clear that it's not a plain function name, and otherwise the excessive punctuation makes some of those invocations hard to read. Discussion: https://www.postgresql.org/message-id/f52c16db-14ed-757d-4b48-7ef360b1631d@2ndquadrant.com
2017-09-07 19:06:23 +03:00
retval = !(tinfo->f_eq(query, key->lower, collation, flinfo) &&
tinfo->f_eq(query, key->upper, collation, flinfo));
Add btree_gist support for searching on "not equals". Jeff Davis, with slight editorialization by me.
2010-08-02 20:26:48 +04:00
break;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
default:
Change TRUE/FALSE to true/false The lower case spellings are C and C++ standard and are used in most parts of the PostgreSQL sources. The upper case spellings are only used in some files/modules. So standardize on the standard spellings. The APIs for ICU, Perl, and Windows define their own TRUE and FALSE, so those are left as is when using those APIs. In code comments, we use the lower-case spelling for the C concepts and keep the upper-case spelling for the SQL concepts. Reviewed-by: Michael Paquier <michael.paquier@gmail.com>
2017-08-16 07:22:32 +03:00
retval = false;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}
Pgindent run for 8.0.
2004-08-29 09:07:03 +04:00
Fix contrib/btree_gist to handle collations properly. Make use of the collation attached to the index column, instead of hard-wiring DEFAULT_COLLATION_OID. (Note: in theory this could require reindexing btree_gist indexes on textual columns, but I rather doubt anyone has one with a non-default declared collation as yet.)
2011-04-23 04:19:58 +04:00
return retval;
New version. Add support for int2, int8, float4, float8, timestamp with/without time zone, time with/without time zone, date, interval, oid, money and macaddr, char, varchar/text, bytea, numeric, bit, varbit, inet/cidr types for GiST
2004-05-28 14:43:32 +04:00
}