Re: Why is citext/regress failing on hamerkop?
Thomas Munro wrote 2024-05-12 06:31: Hamerkop is already green on the 15 and 16 branches, apparently because it's using the pre-meson test stuff that I guess just didn't run the relevant test. In other words, nobody would notice the difference anyway, and a master-only fix would be enough to end this 44-day red streak. Sorry for necroposting, but in our automated testing system we have found some fails of this test. The most recent one was a couple of days ago (see attached files) on PostgreSQL 15.7. Also I've reported this bug some time ago [1], but provided an example only for PostgreSQL 17. Back then the bug was actually found on 15 or 16 branches (no logs remain from couple of months back), but i wanted to show that it was reproducible on 17. I would appreciate if you would backpatch this change to 15 and 16 branches. [1] https://www.postgresql.org/message-id/6885a0b52d06f7e5910d2b6276bbb4e8%40postgrespro.ru Oleg Tselebrovskiy, Postgres ProThe files belonging to this database system will be owned by user "GitLabRunner". This user must also own the server process. The database cluster will be initialized with locale "English_United States.1252". The default text search configuration will be set to "english". Data page checksums are disabled. fixing permissions on existing directory C:/gr-builds/TaKFe3FF/2/pgpro-dev/postgrespro/ci_base ... ok creating subdirectories ... ok selecting dynamic shared memory implementation ... windows selecting default max_connections ... 100 selecting default shared_buffers ... 128MB selecting default time zone ... Europe/Moscow creating configuration files ... ok running bootstrap script ... ok performing post-bootstrap initialization ... ok syncing data to disk ... ok initdb: warning: enabling "trust" authentication for local connections initdb: hint: You can change this by editing pg_hba.conf or using the option -A, or --auth-local and --auth-host, the next time you run initdb. Success. You can now start the database server using: pg_ctl -D ^"C^:^\gr^-builds^\TaKFe3FF^\2^\pgpro^-dev^\postgrespro^\ci^_base^" -l logfile start 2024-07-29 14:41:21.649 MSK [14344] LOG: starting PostgreSQL 15.7, compiled by Visual C++ build 1929, 64-bit 2024-07-29 14:41:21.653 MSK [14344] LOG: listening on IPv6 address "::1", port 5432 2024-07-29 14:41:21.653 MSK [14344] LOG: listening on IPv4 address "127.0.0.1", port 5432 2024-07-29 14:41:21.697 MSK [8052] LOG: database system was shut down at 2024-07-29 14:41:17 MSK 2024-07-29 14:41:21.735 MSK [14344] LOG: database system is ready to accept connections 2024-07-29 14:41:22.859 MSK [9348] ERROR: tablespace location must be an absolute path 2024-07-29 14:41:22.859 MSK [9348] STATEMENT: CREATE TABLESPACE regress_tblspace LOCATION 'relative'; diff -w -U3 C:/gr-builds/TaKFe3FF/2/pgpro-dev/postgrespro/contrib/citext/expected/citext_utf8.out C:/gr-builds/TaKFe3FF/2/pgpro-dev/postgrespro/contrib/citext/results/citext_utf8.out --- C:/gr-builds/TaKFe3FF/2/pgpro-dev/postgrespro/contrib/citext/expected/citext_utf8.out 2024-07-29 13:53:45.259126600 +0300 +++ C:/gr-builds/TaKFe3FF/2/pgpro-dev/postgrespro/contrib/citext/results/citext_utf8.out 2024-07-29 14:43:38.772857200 +0300 @@ -54,7 +54,7 @@ SELECT 'i'::citext = 'Ä°'::citext AS t; t --- - t + f (1 row) -- Regression.
Re: [PROPOSAL] Skip test citext_utf8 on Windows
Michael Paquier писал(а) 2024-03-12 06:24: On Mon, Mar 11, 2024 at 03:21:11PM +0700, Oleg Tselebrovskiy wrote: The proposed patch for skipping test is attached Your attached patch seems to be in binary format. -- Michael Right, I had it saved in not-UTF-8 encoding. Kind of ironic Here's a fixed versiondiff --git a/contrib/citext/expected/citext_utf8.out b/contrib/citext/expected/citext_utf8.out index 5d988dcd485..6c4069f9469 100644 --- a/contrib/citext/expected/citext_utf8.out +++ b/contrib/citext/expected/citext_utf8.out @@ -10,7 +10,8 @@ SELECT getdatabaseencoding() <> 'UTF8' OR (SELECT (datlocprovider = 'c' AND datctype = 'C') OR datlocprovider = 'i' FROM pg_database -WHERE datname=current_database()) +WHERE datname=current_database()) OR + (version() ~ 'windows' OR version() ~ 'Visual C\+\+' OR version() ~ 'mingw32') AS skip_test \gset \if :skip_test \quit diff --git a/contrib/citext/expected/citext_utf8_1.out b/contrib/citext/expected/citext_utf8_1.out index 7065a5da190..d4472b1c36a 100644 --- a/contrib/citext/expected/citext_utf8_1.out +++ b/contrib/citext/expected/citext_utf8_1.out @@ -10,7 +10,8 @@ SELECT getdatabaseencoding() <> 'UTF8' OR (SELECT (datlocprovider = 'c' AND datctype = 'C') OR datlocprovider = 'i' FROM pg_database -WHERE datname=current_database()) +WHERE datname=current_database()) OR + (version() ~ 'windows' OR version() ~ 'Visual C\+\+' OR version() ~ 'mingw32') AS skip_test \gset \if :skip_test \quit diff --git a/contrib/citext/sql/citext_utf8.sql b/contrib/citext/sql/citext_utf8.sql index 34b232d64e2..53775cdcd35 100644 --- a/contrib/citext/sql/citext_utf8.sql +++ b/contrib/citext/sql/citext_utf8.sql @@ -11,7 +11,8 @@ SELECT getdatabaseencoding() <> 'UTF8' OR (SELECT (datlocprovider = 'c' AND datctype = 'C') OR datlocprovider = 'i' FROM pg_database -WHERE datname=current_database()) +WHERE datname=current_database()) OR + (version() ~ 'windows' OR version() ~ 'Visual C\+\+' OR version() ~ 'mingw32') AS skip_test \gset \if :skip_test \quit
[PROPOSAL] Skip test citext_utf8 on Windows
Greetings, everyone!
While running "installchecks" on databases with UTF-8 encoding the test
citext_utf8 fails because of Turkish dotted I like this:
SELECT 'i'::citext = 'İ'::citext AS t;
t
---
- t
+ f
(1 row)
I tried to replicate the test's results by hand and with any collation
that I tried (including --locale="Turkish") this test failed
Also an interesing result of my tesing. If you initialize you DB
with -E utf-8 --locale="Turkish" and then run select LOWER('İ');
the output will be this:
lower
---
İ
(1 row)
Which I find strange since lower() uses collation that was passed
(default in this case but still)
My PostgreSQL version is this:
postgres=# select version();
version
--
PostgreSQL 17devel on x86_64-windows, compiled by gcc-13.1.0, 64-bit
The proposed patch for skipping test is attached
Oleg Tselebrovskiy, Postgres Proÿþd i f f - - g i t a / c o n t r i b / c i t e x t / e x p e c t e d / c i t e x t _ u t f 8 . o u t b / c o n t r i b / c i t e x t / e x p e c t e d / c i t e x t _ u t f 8 . o u t
i n d e x 5 d 9 8 8 d c d 4 8 5 . . 6 c 4 0 6 9 f 9 4 6 9 1 0 0 6 4 4
- - - a / c o n t r i b / c i t e x t / e x p e c t e d / c i t e x t _ u t f 8 . o u t
+ + + b / c o n t r i b / c i t e x t / e x p e c t e d / c i t e x t _ u t f 8 . o u t
@ @ - 1 0 , 7 + 1 0 , 8 @ @
S E L E C T g e t d a t a b a s e e n c o d i n g ( ) < > ' U T F 8 ' O R
( S E L E C T ( d a t l o c p r o v i d e r = ' c ' A N D d a t c t y p e = ' C ' ) O R d a t l o c p r o v i d e r = ' i '
F R O M p g _ d a t a b a s e
- W H E R E d a t n a m e = c u r r e n t _ d a t a b a s e ( ) )
+ W H E R E d a t n a m e = c u r r e n t _ d a t a b a s e ( ) ) O R
+ ( v e r s i o n ( ) ~ ' w i n d o w s ' O R v e r s i o n ( ) ~ ' V i s u a l C \ + \ + ' O R v e r s i o n ( ) ~ ' m i n g w 3 2 ' )
A S s k i p _ t e s t \ g s e t
\ i f : s k i p _ t e s t
\ q u i t
d i f f - - g i t a / c o n t r i b / c i t e x t / e x p e c t e d / c i t e x t _ u t f 8 _ 1 . o u t b / c o n t r i b / c i t e x t / e x p e c t e d / c i t e x t _ u t f 8 _ 1 . o u t
i n d e x 7 0 6 5 a 5 d a 1 9 0 . . d 4 4 7 2 b 1 c 3 6 a 1 0 0 6 4 4
- - - a / c o n t r i b / c i t e x t / e x p e c t e d / c i t e x t _ u t f 8 _ 1 . o u t
+ + + b / c o n t r i b / c i t e x t / e x p e c t e d / c i t e x t _ u t f 8 _ 1 . o u t
@ @ - 1 0 , 7 + 1 0 , 8 @ @
S E L E C T g e t d a t a b a s e e n c o d i n g ( ) < > ' U T F 8 ' O R
( S E L E C T ( d a t l o c p r o v i d e r = ' c ' A N D d a t c t y p e = ' C ' ) O R d a t l o c p r o v i d e r = ' i '
F R O M p g _ d a t a b a s e
- W H E R E d a t n a m e = c u r r e n t _ d a t a b a s e ( ) )
+ W H E R E d a t n a m e = c u r r e n t _ d a t a b a s e ( ) ) O R
+ ( v e r s i o n ( ) ~ ' w i n d o w s ' O R v e r s i o n ( ) ~ ' V i s u a l C \ + \ + ' O R v e r s i o n ( ) ~ ' m i n g w 3 2 ' )
A S s k i p _ t e s t \ g s e t
\ i f : s k i p _ t e s t
\ q u i t
d i f f - - g i t a / c o n t r i b / c i t e x t / s q l / c i t e x t _ u t f 8 . s q l b / c o n t r i b / c i t e x t / s q l / c i t e x t _ u t f 8 . s q l
i n d e x 3 4 b 2 3 2 d 6 4 e 2 . . 5 3 7 7 5 c d c d 3 5 1 0 0 6 4 4
- - - a / c o n t r i b / c i t e x t / s q l / c i t e x t _ u t f 8 . s q l
+ + + b / c o n t r i b / c i t e x t / s q l / c i t e x t _ u t f 8 . s q l
@ @ - 1 1 , 7 + 1 1 , 8 @ @
S E L E C T g e t d a t a b a s e e n c o d i n g ( ) < > ' U T F 8 ' O R
( S E L E C T ( d a t l o c p r o v i d e r = ' c ' A N D d a t c t y p e = ' C ' ) O R d a t l o c p r o v i d e r = ' i '
F R O M p g _ d a t a b a s e
- W H E R E d a t n a m e = c u r r e n t _ d a t a b a s e ( ) )
+ W H E R E d a t n a m e = c u r r e n t _ d a t a b a s e ( ) ) O R
+ ( v e r s i o n ( ) ~ ' w i n d o w s ' O R v e r s i o n ( ) ~ ' V i s u a l C \ + \ + ' O R v e r s i o n ( ) ~ ' m i n g w 3 2 ' )
A S s k i p _ t e s t \ g s e t
\ i f : s k i p _ t e s t
\ q u i t
BRIN integer overflow
Greetings, everyone! While analyzing output of Svace static analyzer [1] I've found a bug Function bringetbitmap that is used in BRIN's IndexAmRoutine should return an int64 value, but the actual return value is int, since totalpages is int and totalpages * 10 is also int. This could lead to integer overflow I suggest to change totalpages to be int64 to avoid potential overflow. Also in all other "amgetbitmap functions" (such as hashgetbitmap, gistgetbitmap, gingetbitmap, blgetbitmap) the return value is of correct int64 type The proposed patch is attached [1] - https://svace.pages.ispras.ru/svace-website/en/ Oleg Tselebrovskiy, Postgres Prodiff --git a/src/backend/access/brin/brin.c b/src/backend/access/brin/brin.c index 1087a9011ea..f72667e484e 100644 --- a/src/backend/access/brin/brin.c +++ b/src/backend/access/brin/brin.c @@ -559,7 +559,7 @@ bringetbitmap(IndexScanDesc scan, TIDBitmap *tbm) BrinOpaque *opaque; BlockNumber nblocks; BlockNumber heapBlk; - int totalpages = 0; + int64 totalpages = 0; FmgrInfo *consistentFn; MemoryContext oldcxt; MemoryContext perRangeCxt;
Re: Returning non-terminated string in ECPG Informix-compatible function
Thanks for review!
I added a regression test that is based on code from previous email
New patch is attached
Oleg Tselebrovskiy, Postgres Prodiff --git a/src/interfaces/ecpg/compatlib/informix.c b/src/interfaces/ecpg/compatlib/informix.c
index dccf39582da..80d40aa3e09 100644
--- a/src/interfaces/ecpg/compatlib/informix.c
+++ b/src/interfaces/ecpg/compatlib/informix.c
@@ -654,7 +654,7 @@ intoasc(interval * i, char *str)
if (!tmp)
return -errno;
- memcpy(str, tmp, strlen(tmp));
+ strcpy(str, tmp);
free(tmp);
return 0;
}
diff --git a/src/interfaces/ecpg/pgtypeslib/dt_common.c b/src/interfaces/ecpg/pgtypeslib/dt_common.c
index 99bdc94d6d7..d4ca0cbff6e 100644
--- a/src/interfaces/ecpg/pgtypeslib/dt_common.c
+++ b/src/interfaces/ecpg/pgtypeslib/dt_common.c
@@ -2659,6 +2659,8 @@ PGTYPEStimestamp_defmt_scan(char **str, char *fmt, timestamp * d,
*/
pfmt++;
tmp = pgtypes_alloc(strlen("%m/%d/%y") + strlen(pstr) + 1);
+if(!tmp)
+ return 1;
strcpy(tmp, "%m/%d/%y");
strcat(tmp, pfmt);
err = PGTYPEStimestamp_defmt_scan(&pstr, tmp, d, year, month, day, hour, minute, second, tz);
@@ -2784,6 +2786,8 @@ PGTYPEStimestamp_defmt_scan(char **str, char *fmt, timestamp * d,
case 'r':
pfmt++;
tmp = pgtypes_alloc(strlen("%I:%M:%S %p") + strlen(pstr) + 1);
+if(!tmp)
+ return 1;
strcpy(tmp, "%I:%M:%S %p");
strcat(tmp, pfmt);
err = PGTYPEStimestamp_defmt_scan(&pstr, tmp, d, year, month, day, hour, minute, second, tz);
@@ -2792,6 +2796,8 @@ PGTYPEStimestamp_defmt_scan(char **str, char *fmt, timestamp * d,
case 'R':
pfmt++;
tmp = pgtypes_alloc(strlen("%H:%M") + strlen(pstr) + 1);
+if(!tmp)
+ return 1;
strcpy(tmp, "%H:%M");
strcat(tmp, pfmt);
err = PGTYPEStimestamp_defmt_scan(&pstr, tmp, d, year, month, day, hour, minute, second, tz);
@@ -2837,6 +2843,8 @@ PGTYPEStimestamp_defmt_scan(char **str, char *fmt, timestamp * d,
case 'T':
pfmt++;
tmp = pgtypes_alloc(strlen("%H:%M:%S") + strlen(pstr) + 1);
+if(!tmp)
+ return 1;
strcpy(tmp, "%H:%M:%S");
strcat(tmp, pfmt);
err = PGTYPEStimestamp_defmt_scan(&pstr, tmp, d, year, month, day, hour, minute, second, tz);
diff --git a/src/interfaces/ecpg/test/compat_informix/.gitignore b/src/interfaces/ecpg/test/compat_informix/.gitignore
index f97706ba4be..6967ae77cd2 100644
--- a/src/interfaces/ecpg/test/compat_informix/.gitignore
+++ b/src/interfaces/ecpg/test/compat_informix/.gitignore
@@ -4,6 +4,8 @@
/dec_test.c
/describe
/describe.c
+/intoasc
+/intoasc.c
/rfmtdate
/rfmtdate.c
/rfmtlong
diff --git a/src/interfaces/ecpg/test/compat_informix/Makefile b/src/interfaces/ecpg/test/compat_informix/Makefile
index d50fdc29fd1..638b4e0af78 100644
--- a/src/interfaces/ecpg/test/compat_informix/Makefile
+++ b/src/interfaces/ecpg/test/compat_informix/Makefile
@@ -16,7 +16,8 @@ TESTS = test_informix test_informix.c \
rnull rnull.c \
sqlda sqlda.c \
describe describe.c \
-charfuncs charfuncs.c
+charfuncs charfuncs.c \
+intoasc intoasc.c
all: $(TESTS)
diff --git a/src/interfaces/ecpg/test/compat_informix/intoasc.pgc b/src/interfaces/ecpg/test/compat_informix/intoasc.pgc
new file mode 100644
index 000..d13c83bb7a7
--- /dev/null
+++ b/src/interfaces/ecpg/test/compat_informix/intoasc.pgc
@@ -0,0 +1,21 @@
+#include
+#include
+
+#include "pgtypes_interval.h"
+
+EXEC SQL BEGIN DECLARE SECTION;
+char dirty_str[100] = "a__c_d_";
+interval *interval_ptr;
+EXEC SQL END DECLARE SECTION;
+
+int main()
+{
+interval_ptr = (interval *) malloc(sizeof(interval));
+interval_ptr->time = 1;
+interval_ptr->month = 240;
+
+printf("dirty_str contents before intoasc: %s\n", dirty_str);
+intoasc(interval_ptr, dirty_str);
+printf("dirty_str contents after intoasc: %s\n", dirty_str);
+return 0;
+}
diff --git a/src/interfaces/ecpg/test/compat_informix/meson.build b/src/interfaces/ecpg/test/compat_informix/meson.build
index e2f8802330d..7e4790933ad 100644
--- a/src/interfaces/ecpg/test/compat_informix/meson.build
+++ b/src/interfaces/ecpg/test/compat_informix/meson.build
@@ -4,6 +4,7 @@ pgc_files = [
'charfuncs',
'dec_test',
'describe',
+ 'intoasc',
'rfmtdate',
'rfmtlong',
'rnull',
diff --git a/src/interfaces/ecpg/test/ecpg_schedule b/src/interfaces/ecpg/test/ecpg_schedule
index 39814a39c17..f9c0a0e3c00 100644
--- a/src/interfaces/ecpg/test/ecpg_schedule
+++ b/src/interfaces/ecpg/test/ecpg_schedule
@@ -7,6 +7,7 @@ test: compat_informix/sqlda
test: compat_informix/describe
test: compat_informix/test_informix
test: compat_informix/test_informix2
+test: compat_informix/i
xmlBufferCreate return value not checked in pgxmlNodeSetToText
Greetings, everyone!
While analyzing output of Svace static analyzer [1] I've found a bug.
In function pgxmlNodeSetToText there is a call of xmlBufferCreate that
doesn't
have its return value checked. In all four other calls of
xmlBufferCreate there
is a try...catch that checks the return value inside.
I suggest to add the same checks here that are used in other four calls
of
xmlBufferCreate.
The proposed patch is attached.
[1] - https://svace.pages.ispras.ru/svace-website/en/
Oleg Tselebrovskiy, Postgres Prodiff --git a/contrib/xml2/xpath.c b/contrib/xml2/xpath.c
index 94641930f7b..2d72ade9c20 100644
--- a/contrib/xml2/xpath.c
+++ b/contrib/xml2/xpath.c
@@ -122,62 +122,85 @@ pgxmlNodeSetToText(xmlNodeSetPtr nodeset,
xmlChar *septagname,
xmlChar *plainsep)
{
- xmlBufferPtr buf;
+ xmlBufferPtr buf = NULL;
xmlChar*result;
int i;
+ PgXmlErrorContext *xmlerrcxt;
- buf = xmlBufferCreate();
+ xmlerrcxt = pg_xml_init(PG_XML_STRICTNESS_LEGACY);
- if ((toptagname != NULL) && (xmlStrlen(toptagname) > 0))
- {
- xmlBufferWriteChar(buf, "<");
- xmlBufferWriteCHAR(buf, toptagname);
- xmlBufferWriteChar(buf, ">");
- }
- if (nodeset != NULL)
+ PG_TRY();
{
- for (i = 0; i < nodeset->nodeNr; i++)
- {
- if (plainsep != NULL)
- {
-xmlBufferWriteCHAR(buf,
- xmlXPathCastNodeToString(nodeset->nodeTab[i]));
+ buf = xmlBufferCreate();
-/* If this isn't the last entry, write the plain sep. */
-if (i < (nodeset->nodeNr) - 1)
- xmlBufferWriteChar(buf, (char *) plainsep);
- }
- else
+ if (buf == NULL)
+ xml_ereport(xmlerrcxt, ERROR, ERRCODE_OUT_OF_MEMORY,
+ "could not allocate xmlBuffer");
+
+ if ((toptagname != NULL) && (xmlStrlen(toptagname) > 0))
+ {
+ xmlBufferWriteChar(buf, "<");
+ xmlBufferWriteCHAR(buf, toptagname);
+ xmlBufferWriteChar(buf, ">");
+ }
+ if (nodeset != NULL)
+ {
+ for (i = 0; i < nodeset->nodeNr; i++)
{
-if ((septagname != NULL) && (xmlStrlen(septagname) > 0))
+if (plainsep != NULL)
{
- xmlBufferWriteChar(buf, "<");
- xmlBufferWriteCHAR(buf, septagname);
- xmlBufferWriteChar(buf, ">");
-}
-xmlNodeDump(buf,
- nodeset->nodeTab[i]->doc,
- nodeset->nodeTab[i],
- 1, 0);
+ xmlBufferWriteCHAR(buf,
+ xmlXPathCastNodeToString(nodeset->nodeTab[i]));
-if ((septagname != NULL) && (xmlStrlen(septagname) > 0))
+ /* If this isn't the last entry, write the plain sep. */
+ if (i < (nodeset->nodeNr) - 1)
+ xmlBufferWriteChar(buf, (char *) plainsep);
+}
+else
{
- xmlBufferWriteChar(buf, "");
+ if ((septagname != NULL) && (xmlStrlen(septagname) > 0))
+ {
+ xmlBufferWriteChar(buf, "<");
+ xmlBufferWriteCHAR(buf, septagname);
+ xmlBufferWriteChar(buf, ">");
+ }
+ xmlNodeDump(buf,
+nodeset->nodeTab[i]->doc,
+nodeset->nodeTab[i],
+1, 0);
+
+ if ((septagname != NULL) && (xmlStrlen(septagname) > 0))
+ {
+ xmlBufferWriteChar(buf, "");
+ }
}
}
}
- }
- if ((toptagname != NULL) && (xmlStrlen(toptagname) > 0))
+ if ((toptagname != NULL) && (xmlStrlen(toptagname) > 0))
+ {
+ xmlBufferWriteChar(buf, "");
+ }
+ result = xmlStrdup(buf->content);
+ }
+ PG_CATCH();
{
- xmlBufferWriteChar(buf, "");
+ if (buf)
+ xmlBufferFree(buf);
+
+ pg_xml_done(xmlerrcxt, true);
+
+ PG_RE_THROW();
}
- result = xmlStrdup(buf->content);
+ PG_END_TRY();
+
xmlBufferFree(buf);
+ pg_xml_done(xmlerrcxt, false);
+
return result;
}
Re: Returning non-terminated string in ECPG Informix-compatible function
Greetings again.
I was looking through more static analyzer output and found another
problem.
In ecpg/pgtypeslib/dt_common.c there are 4 calls of pgtypes_alloc.
This function uses calloc and returns NULL if OOM, but we don't check
its
return value and immediately pass it to strcpy, which could lead to
segfault.
I suggest adding a check for a return value since all other calls of
pgtypes_alloc are checked for NULL.
A proposed patch (with previous and current changes) is attached
Oleg Tselebrovskiy, Postgres Prodiff --git a/src/interfaces/ecpg/compatlib/informix.c b/src/interfaces/ecpg/compatlib/informix.c
index dccf39582da..80d40aa3e09 100644
--- a/src/interfaces/ecpg/compatlib/informix.c
+++ b/src/interfaces/ecpg/compatlib/informix.c
@@ -654,7 +654,7 @@ intoasc(interval * i, char *str)
if (!tmp)
return -errno;
- memcpy(str, tmp, strlen(tmp));
+ strcpy(str, tmp);
free(tmp);
return 0;
}
diff --git a/src/interfaces/ecpg/pgtypeslib/dt_common.c b/src/interfaces/ecpg/pgtypeslib/dt_common.c
index 99bdc94d6d7..d4ca0cbff6e 100644
--- a/src/interfaces/ecpg/pgtypeslib/dt_common.c
+++ b/src/interfaces/ecpg/pgtypeslib/dt_common.c
@@ -2659,6 +2659,8 @@ PGTYPEStimestamp_defmt_scan(char **str, char *fmt, timestamp * d,
*/
pfmt++;
tmp = pgtypes_alloc(strlen("%m/%d/%y") + strlen(pstr) + 1);
+if(!tmp)
+ return 1;
strcpy(tmp, "%m/%d/%y");
strcat(tmp, pfmt);
err = PGTYPEStimestamp_defmt_scan(&pstr, tmp, d, year, month, day, hour, minute, second, tz);
@@ -2784,6 +2786,8 @@ PGTYPEStimestamp_defmt_scan(char **str, char *fmt, timestamp * d,
case 'r':
pfmt++;
tmp = pgtypes_alloc(strlen("%I:%M:%S %p") + strlen(pstr) + 1);
+if(!tmp)
+ return 1;
strcpy(tmp, "%I:%M:%S %p");
strcat(tmp, pfmt);
err = PGTYPEStimestamp_defmt_scan(&pstr, tmp, d, year, month, day, hour, minute, second, tz);
@@ -2792,6 +2796,8 @@ PGTYPEStimestamp_defmt_scan(char **str, char *fmt, timestamp * d,
case 'R':
pfmt++;
tmp = pgtypes_alloc(strlen("%H:%M") + strlen(pstr) + 1);
+if(!tmp)
+ return 1;
strcpy(tmp, "%H:%M");
strcat(tmp, pfmt);
err = PGTYPEStimestamp_defmt_scan(&pstr, tmp, d, year, month, day, hour, minute, second, tz);
@@ -2837,6 +2843,8 @@ PGTYPEStimestamp_defmt_scan(char **str, char *fmt, timestamp * d,
case 'T':
pfmt++;
tmp = pgtypes_alloc(strlen("%H:%M:%S") + strlen(pstr) + 1);
+if(!tmp)
+ return 1;
strcpy(tmp, "%H:%M:%S");
strcat(tmp, pfmt);
err = PGTYPEStimestamp_defmt_scan(&pstr, tmp, d, year, month, day, hour, minute, second, tz);
Re: Returning non-terminated string in ECPG Informix-compatible function
Here's the code for bug reproduction:
#include
#include
EXEC SQL INCLUDE pgtypes_interval.h;
EXEC SQL INCLUDE ecpg_informix.h;
EXEC SQL BEGIN DECLARE SECTION;
char dirty_str[100] = "a__c_d_";
interval *interval_ptr;
EXEC SQL END DECLARE SECTION;
int main()
{
interval_ptr = (interval *) malloc(sizeof(interval));
interval_ptr->time = 1;
interval_ptr->month = 240;
printf("dirty_str contents before intoasc: %s\n", dirty_str);
intoasc(interval_ptr, dirty_str);
printf("dirty_str contents after intoasc: %s\n", dirty_str);
return 0;
}
And here's the output:
dirty_str contents before intoasc:
a__c_d_
dirty_str contents after intoasc: @ 20 years 1 min 40
secs_d_
I compiled it with following commands (provided for quicker
reproduction):
/path/to/pgsql/bin/ecpg informix_bug_example.pgc
gcc -I/path/to/pgsql/include -c informix_bug_example.c
gcc -o informix_bug_example informix_bug_example.o -L/path/to/pgsql/lib
-lecpg -lecpg_compat
I've also found at least one project that uses intoasc() in it -
https://github.com/credativ/informix_fdw/
Oleg Tselebrovskiy, Postgres Pro
