Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commitc8c7c93

Browse files
committed
Fix PL/Tcl's encoding conversion logic.
PL/Tcl appears to contain logic to convert strings between the databaseencoding and UTF8, which is the only encoding modern Tcl will deal with.However, that code has been disabled since commit0348951, whichmade it "#if defined(UNICODE_CONVERSION)" and neglected to provide any wayfor that symbol to become defined. That might have been all right backin 2001, but these days we take a dim view of allowing strings withincorrect encoding into the database.Remove the conditional compilation, fix warnings about signed/unsigned charconversions, clean up assorted places that didn't bother with conversions.(Notably, there were lots of assumptions that database table and fieldnames didn't need conversion...)Add a regression test based on plpython_unicode. It's not terriblythorough, but better than no test at all.
1 parente260932 commitc8c7c93

File tree

4 files changed

+145
-72
lines changed

4 files changed

+145
-72
lines changed

‎src/pl/tcl/Makefile‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -29,7 +29,7 @@ DATA = pltcl.control pltcl--1.0.sql pltcl--unpackaged--1.0.sql \
2929
pltclu.control pltclu--1.0.sql pltclu--unpackaged--1.0.sql
3030

3131
REGRESS_OPTS = --dbname=$(PL_TESTDB) --load-extension=pltcl
32-
REGRESS = pltcl_setup pltcl_queries
32+
REGRESS = pltcl_setup pltcl_queries pltcl_unicode
3333

3434
# Tcl on win32 ships with import libraries only for Microsoft Visual C++,
3535
# which are not compatible with mingw gcc. Therefore we need to build a
Lines changed: 45 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,45 @@
1+
--
2+
-- Unicode handling
3+
--
4+
-- Note: this test case is known to fail if the database encoding is
5+
-- EUC_CN, EUC_JP, EUC_KR, or EUC_TW, for lack of any equivalent to
6+
-- U+00A0 (no-break space) in those encodings. However, testing with
7+
-- plain ASCII data would be rather useless, so we must live with that.
8+
--
9+
SET client_encoding TO UTF8;
10+
CREATE TABLE unicode_test (
11+
testvalue text NOT NULL
12+
);
13+
CREATE FUNCTION unicode_return() RETURNS text AS $$
14+
return "\xA0"
15+
$$ LANGUAGE pltcl;
16+
CREATE FUNCTION unicode_trigger() RETURNS trigger AS $$
17+
set NEW(testvalue) "\xA0"
18+
return [array get NEW]
19+
$$ LANGUAGE pltcl;
20+
CREATE TRIGGER unicode_test_bi BEFORE INSERT ON unicode_test
21+
FOR EACH ROW EXECUTE PROCEDURE unicode_trigger();
22+
CREATE FUNCTION unicode_plan1() RETURNS text AS $$
23+
set plan [ spi_prepare {SELECT $1 AS testvalue} [ list "text" ] ]
24+
spi_execp $plan [ list "\xA0" ]
25+
return $testvalue
26+
$$ LANGUAGE pltcl;
27+
SELECT unicode_return();
28+
unicode_return
29+
----------------
30+
 
31+
(1 row)
32+
33+
INSERT INTO unicode_test (testvalue) VALUES ('test');
34+
SELECT * FROM unicode_test;
35+
testvalue
36+
-----------
37+
 
38+
(1 row)
39+
40+
SELECT unicode_plan1();
41+
unicode_plan1
42+
---------------
43+
 
44+
(1 row)
45+

‎src/pl/tcl/pltcl.c‎

Lines changed: 61 additions & 71 deletions
Original file line numberDiff line numberDiff line change
@@ -21,6 +21,7 @@
2121
#include"commands/trigger.h"
2222
#include"executor/spi.h"
2323
#include"fmgr.h"
24+
#include"mb/pg_wchar.h"
2425
#include"miscadmin.h"
2526
#include"nodes/makefuncs.h"
2627
#include"parser/parse_type.h"
@@ -33,6 +34,8 @@
3334
#include"utils/typcache.h"
3435

3536

37+
PG_MODULE_MAGIC;
38+
3639
#defineHAVE_TCL_VERSION(maj,min) \
3740
((TCL_MAJOR_VERSION > maj) || \
3841
(TCL_MAJOR_VERSION == maj && TCL_MINOR_VERSION >= min))
@@ -51,43 +54,44 @@
5154
#undef TEXTDOMAIN
5255
#defineTEXTDOMAIN PG_TEXTDOMAIN("pltcl")
5356

54-
#if defined(UNICODE_CONVERSION)
5557

56-
#include"mb/pg_wchar.h"
58+
/*
59+
* Support for converting between UTF8 (which is what all strings going into
60+
* or out of Tcl should be) and the database encoding.
61+
*
62+
* If you just use utf_u2e() or utf_e2u() directly, they will leak some
63+
* palloc'd space when doing a conversion. This is not worth worrying about
64+
* if it only happens, say, once per PL/Tcl function call. If it does seem
65+
* worth worrying about, use the wrapper macros.
66+
*/
5767

58-
staticunsignedchar*
59-
utf_u2e(unsignedchar*src)
68+
staticinlinechar*
69+
utf_u2e(constchar*src)
6070
{
61-
return (unsignedchar*)pg_any_to_server((char*)src,
62-
strlen(src),
63-
PG_UTF8);
71+
returnpg_any_to_server(src,strlen(src),PG_UTF8);
6472
}
6573

66-
staticunsignedchar*
67-
utf_e2u(unsignedchar*src)
74+
staticinlinechar*
75+
utf_e2u(constchar*src)
6876
{
69-
return (unsignedchar*)pg_server_to_any((char*)src,
70-
strlen(src),
71-
PG_UTF8);
77+
returnpg_server_to_any(src,strlen(src),PG_UTF8);
7278
}
7379

74-
#definePLTCL_UTF
75-
#defineUTF_BEGIN do { \
76-
unsigned char *_pltcl_utf_src; \
77-
unsigned char *_pltcl_utf_dst
78-
#defineUTF_END if (_pltcl_utf_src!=_pltcl_utf_dst) \
79-
pfree(_pltcl_utf_dst); } while (0)
80-
#defineUTF_U2E(x) (_pltcl_utf_dst=utf_u2e(_pltcl_utf_src=(x)))
81-
#defineUTF_E2U(x) (_pltcl_utf_dst=utf_e2u(_pltcl_utf_src=(x)))
82-
#else/* !PLTCL_UTF */
83-
84-
#defineUTF_BEGIN
85-
#defineUTF_END
86-
#defineUTF_U2E(x) (x)
87-
#defineUTF_E2U(x) (x)
88-
#endif/* PLTCL_UTF */
80+
#defineUTF_BEGIN \
81+
do { \
82+
const char *_pltcl_utf_src = NULL; \
83+
char *_pltcl_utf_dst = NULL
8984

90-
PG_MODULE_MAGIC;
85+
#defineUTF_END \
86+
if (_pltcl_utf_src != (const char *) _pltcl_utf_dst) \
87+
pfree(_pltcl_utf_dst); \
88+
} while (0)
89+
90+
#defineUTF_U2E(x) \
91+
(_pltcl_utf_dst = utf_u2e(_pltcl_utf_src = (x)))
92+
93+
#defineUTF_E2U(x) \
94+
(_pltcl_utf_dst = utf_e2u(_pltcl_utf_src = (x)))
9195

9296

9397
/**********************************************************************
@@ -572,14 +576,10 @@ pltcl_init_load_unknown(Tcl_Interp *interp)
572576
SPI_freetuptable(SPI_tuptable);
573577

574578
if (tcl_rc!=TCL_OK)
575-
{
576-
UTF_BEGIN;
577579
ereport(ERROR,
578580
(errcode(ERRCODE_EXTERNAL_ROUTINE_EXCEPTION),
579581
errmsg("could not load module \"unknown\": %s",
580-
UTF_U2E(Tcl_GetStringResult(interp)))));
581-
UTF_END;
582-
}
582+
utf_u2e(Tcl_GetStringResult(interp)))));
583583

584584
relation_close(pmrel,AccessShareLock);
585585
}
@@ -804,14 +804,10 @@ pltcl_func_handler(PG_FUNCTION_ARGS, bool pltrusted)
804804
prodesc->result_typioparam,
805805
-1);
806806
else
807-
{
808-
UTF_BEGIN;
809807
retval=InputFunctionCall(&prodesc->result_in_func,
810-
UTF_U2E((char*)Tcl_GetStringResult(interp)),
808+
utf_u2e(Tcl_GetStringResult(interp)),
811809
prodesc->result_typioparam,
812810
-1);
813-
UTF_END;
814-
}
815811

816812
returnretval;
817813
}
@@ -866,13 +862,13 @@ pltcl_trigger_handler(PG_FUNCTION_ARGS, bool pltrusted)
866862

867863
PG_TRY();
868864
{
869-
/* The procedure name */
865+
/* The procedure name(note this is all ASCII, so no utf_e2u)*/
870866
Tcl_ListObjAppendElement(NULL,tcl_cmd,
871867
Tcl_NewStringObj(prodesc->internal_proname,-1));
872868

873869
/* The trigger name for argument TG_name */
874870
Tcl_ListObjAppendElement(NULL,tcl_cmd,
875-
Tcl_NewStringObj(trigdata->tg_trigger->tgname,-1));
871+
Tcl_NewStringObj(utf_e2u(trigdata->tg_trigger->tgname),-1));
876872

877873
/* The oid of the trigger relation for argument TG_relid */
878874
/* Consider not converting to a string for more performance? */
@@ -885,13 +881,13 @@ pltcl_trigger_handler(PG_FUNCTION_ARGS, bool pltrusted)
885881
/* The name of the table the trigger is acting on: TG_table_name */
886882
stroid=SPI_getrelname(trigdata->tg_relation);
887883
Tcl_ListObjAppendElement(NULL,tcl_cmd,
888-
Tcl_NewStringObj(stroid,-1));
884+
Tcl_NewStringObj(utf_e2u(stroid),-1));
889885
pfree(stroid);
890886

891887
/* The schema of the table the trigger is acting on: TG_table_schema */
892888
stroid=SPI_getnspname(trigdata->tg_relation);
893889
Tcl_ListObjAppendElement(NULL,tcl_cmd,
894-
Tcl_NewStringObj(stroid,-1));
890+
Tcl_NewStringObj(utf_e2u(stroid),-1));
895891
pfree(stroid);
896892

897893
/* A list of attribute names for argument TG_relatts */
@@ -903,7 +899,7 @@ pltcl_trigger_handler(PG_FUNCTION_ARGS, bool pltrusted)
903899
Tcl_ListObjAppendElement(NULL,tcl_trigtup,Tcl_NewObj());
904900
else
905901
Tcl_ListObjAppendElement(NULL,tcl_trigtup,
906-
Tcl_NewStringObj(NameStr(tupdesc->attrs[i]->attname),-1));
902+
Tcl_NewStringObj(utf_e2u(NameStr(tupdesc->attrs[i]->attname)),-1));
907903
}
908904
Tcl_ListObjAppendElement(NULL,tcl_cmd,tcl_trigtup);
909905

@@ -1001,7 +997,7 @@ pltcl_trigger_handler(PG_FUNCTION_ARGS, bool pltrusted)
1001997
/* Finally append the arguments from CREATE TRIGGER */
1002998
for (i=0;i<trigdata->tg_trigger->tgnargs;i++)
1003999
Tcl_ListObjAppendElement(NULL,tcl_cmd,
1004-
Tcl_NewStringObj(trigdata->tg_trigger->tgargs[i],-1));
1000+
Tcl_NewStringObj(utf_e2u(trigdata->tg_trigger->tgargs[i]),-1));
10051001

10061002
}
10071003
PG_CATCH();
@@ -1048,14 +1044,10 @@ pltcl_trigger_handler(PG_FUNCTION_ARGS, bool pltrusted)
10481044
************************************************************/
10491045
if (Tcl_SplitList(interp,result,
10501046
&ret_numvals,&ret_values)!=TCL_OK)
1051-
{
1052-
UTF_BEGIN;
10531047
ereport(ERROR,
10541048
(errcode(ERRCODE_E_R_I_E_TRIGGER_PROTOCOL_VIOLATED),
10551049
errmsg("could not split return value from trigger: %s",
1056-
UTF_U2E(Tcl_GetStringResult(interp)))));
1057-
UTF_END;
1058-
}
1050+
utf_u2e(Tcl_GetStringResult(interp)))));
10591051

10601052
/* Use a TRY to ensure ret_values will get freed */
10611053
PG_TRY();
@@ -1078,8 +1070,8 @@ pltcl_trigger_handler(PG_FUNCTION_ARGS, bool pltrusted)
10781070

10791071
for (i=0;i<ret_numvals;i+=2)
10801072
{
1081-
constchar*ret_name=ret_values[i];
1082-
constchar*ret_value=ret_values[i+1];
1073+
char*ret_name=utf_u2e(ret_values[i]);
1074+
char*ret_value=utf_u2e(ret_values[i+1]);
10831075
intattnum;
10841076
Oidtypinput;
10851077
Oidtypioparam;
@@ -1123,13 +1115,11 @@ pltcl_trigger_handler(PG_FUNCTION_ARGS, bool pltrusted)
11231115
/************************************************************
11241116
* Set the attribute to NOT NULL and convert the contents
11251117
************************************************************/
1126-
modnulls[attnum-1]=' ';
1127-
UTF_BEGIN;
11281118
modvalues[attnum-1]=InputFunctionCall(&finfo,
1129-
(char*)UTF_U2E(ret_value),
1119+
ret_value,
11301120
typioparam,
11311121
tupdesc->attrs[attnum-1]->atttypmod);
1132-
UTF_END;
1122+
modnulls[attnum-1]=' ';
11331123
}
11341124

11351125
rettup=SPI_modifytuple(trigdata->tg_relation,rettup,tupdesc->natts,
@@ -1183,9 +1173,9 @@ pltcl_event_trigger_handler(PG_FUNCTION_ARGS, bool pltrusted)
11831173
Tcl_ListObjAppendElement(NULL,tcl_cmd,
11841174
Tcl_NewStringObj(prodesc->internal_proname,-1));
11851175
Tcl_ListObjAppendElement(NULL,tcl_cmd,
1186-
Tcl_NewStringObj(tdata->event,-1));
1176+
Tcl_NewStringObj(utf_e2u(tdata->event),-1));
11871177
Tcl_ListObjAppendElement(NULL,tcl_cmd,
1188-
Tcl_NewStringObj(tdata->tag,-1));
1178+
Tcl_NewStringObj(utf_e2u(tdata->tag),-1));
11891179

11901180
tcl_rc=Tcl_EvalObjEx(interp,tcl_cmd, (TCL_EVAL_DIRECT |TCL_EVAL_GLOBAL));
11911181

@@ -1217,18 +1207,13 @@ throw_tcl_error(Tcl_Interp *interp, const char *proname)
12171207
char*emsg;
12181208
char*econtext;
12191209

1220-
UTF_BEGIN;
1221-
emsg=pstrdup(UTF_U2E(Tcl_GetStringResult(interp)));
1222-
UTF_END;
1223-
UTF_BEGIN;
1224-
econtext=UTF_U2E((char*)Tcl_GetVar(interp,"errorInfo",
1225-
TCL_GLOBAL_ONLY));
1210+
emsg=pstrdup(utf_u2e(Tcl_GetStringResult(interp)));
1211+
econtext=utf_u2e(Tcl_GetVar(interp,"errorInfo",TCL_GLOBAL_ONLY));
12261212
ereport(ERROR,
12271213
(errcode(ERRCODE_EXTERNAL_ROUTINE_EXCEPTION),
12281214
errmsg("%s",emsg),
12291215
errcontext("%s\nin PL/Tcl function \"%s\"",
12301216
econtext,proname)));
1231-
UTF_END;
12321217
}
12331218

12341219

@@ -1315,7 +1300,7 @@ compile_pltcl_function(Oid fn_oid, Oid tgreloid,
13151300
/************************************************************
13161301
* Build our internal proc name from the function's Oid. Append
13171302
* "_trigger" when appropriate to ensure the normal and trigger
1318-
* cases are kept separate.
1303+
* cases are kept separate. Note name must be all-ASCII.
13191304
************************************************************/
13201305
if (!is_trigger&& !is_event_trigger)
13211306
snprintf(internal_proname,sizeof(internal_proname),
@@ -1570,13 +1555,11 @@ compile_pltcl_function(Oid fn_oid, Oid tgreloid,
15701555
free(prodesc->user_proname);
15711556
free(prodesc->internal_proname);
15721557
free(prodesc);
1573-
UTF_BEGIN;
15741558
ereport(ERROR,
15751559
(errcode(ERRCODE_EXTERNAL_ROUTINE_EXCEPTION),
15761560
errmsg("could not create internal procedure \"%s\": %s",
15771561
internal_proname,
1578-
UTF_U2E(Tcl_GetStringResult(interp)))));
1579-
UTF_END;
1562+
utf_u2e(Tcl_GetStringResult(interp)))));
15801563
}
15811564

15821565
/************************************************************
@@ -2212,7 +2195,8 @@ pltcl_SPI_prepare(ClientData cdata, Tcl_Interp *interp,
22122195
* Prepare the plan and check for errors
22132196
************************************************************/
22142197
UTF_BEGIN;
2215-
qdesc->plan=SPI_prepare(UTF_U2E(Tcl_GetString(objv[1])),nargs,qdesc->argtypes);
2198+
qdesc->plan=SPI_prepare(UTF_U2E(Tcl_GetString(objv[1])),
2199+
nargs,qdesc->argtypes);
22162200
UTF_END;
22172201

22182202
if (qdesc->plan==NULL)
@@ -2434,7 +2418,7 @@ pltcl_SPI_execute_plan(ClientData cdata, Tcl_Interp *interp,
24342418
{
24352419
UTF_BEGIN;
24362420
argvalues[j]=InputFunctionCall(&qdesc->arginfuncs[j],
2437-
(char*)UTF_U2E(Tcl_GetString(callObjv[j])),
2421+
UTF_U2E(Tcl_GetString(callObjv[j])),
24382422
qdesc->argtypioparams[j],
24392423
-1);
24402424
UTF_END;
@@ -2483,6 +2467,8 @@ pltcl_SPI_lastoid(ClientData cdata, Tcl_Interp *interp,
24832467
/**********************************************************************
24842468
* pltcl_set_tuple_values() - Set variables for all attributes
24852469
* of a given tuple
2470+
*
2471+
* Note: arrayname is presumed to be UTF8; it usually came from Tcl
24862472
**********************************************************************/
24872473
staticvoid
24882474
pltcl_set_tuple_values(Tcl_Interp*interp,constchar*arrayname,
@@ -2524,7 +2510,9 @@ pltcl_set_tuple_values(Tcl_Interp *interp, const char *arrayname,
25242510
/************************************************************
25252511
* Get the attribute name
25262512
************************************************************/
2527-
attname=NameStr(tupdesc->attrs[i]->attname);
2513+
UTF_BEGIN;
2514+
attname=pstrdup(UTF_E2U(NameStr(tupdesc->attrs[i]->attname)));
2515+
UTF_END;
25282516

25292517
/************************************************************
25302518
* Get the attributes value
@@ -2552,6 +2540,8 @@ pltcl_set_tuple_values(Tcl_Interp *interp, const char *arrayname,
25522540
}
25532541
else
25542542
Tcl_UnsetVar2(interp,*arrptr,*nameptr,0);
2543+
2544+
pfree((char*)attname);
25552545
}
25562546
}
25572547

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp