Adjust comments about avoiding use of printf's %.*s.

My initial impression that glibc was measuring the precision in characters
(which is what the Linux man page says it does) was incorrect.  It does take
the precision to be in bytes, but it also tries to truncate the string at a
character boundary.  The bottom line remains the same: it will mess up
if the string is not in the encoding it expects, so we need to avoid %.*s
anytime there's a significant risk of that.  Previous code changes are still
good, but adjust the comments to reflect this knowledge.  Per research by
Hernan Gonzalez.

diff --git a/src/backend/parser/scansup.c b/src/backend/parser/scansup.c
index 417c79dd14ea29e02236f1f221ad4f92709e300e..94082f77a042570ac999580c49a70f456827d2cf 100644 (file)
--- a/src/backend/parser/scansup.c
+++ b/src/backend/parser/scansup.c
@@ -9,7 +9,7 @@
  *
  *
  * IDENTIFICATION
- *       $PostgreSQL: pgsql/src/backend/parser/scansup.c,v 1.40 2010/05/08 16:39:49 tgl Exp $
+ *       $PostgreSQL: pgsql/src/backend/parser/scansup.c,v 1.41 2010/05/09 02:15:59 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -178,8 +178,8 @@ truncate_identifier(char *ident, int len, bool warn)
                if (warn)
                {
                        /*
-                        * Cannot use %.*s here because some machines interpret %s's
-                        * precision in characters, others in bytes.
+                        * We avoid using %.*s here because it can misbehave if the data
+                        * is not valid in what libc thinks is the prevailing encoding.
                         */
                        char    buf[NAMEDATALEN];
 

diff --git a/src/backend/tsearch/wparser_def.c b/src/backend/tsearch/wparser_def.c
index d2e47ceaf5e32dd53f53014055ffb26b67d45dab..cda28103c24cb89aa78b9bebe893ff33aecb5167 100644 (file)
--- a/src/backend/tsearch/wparser_def.c
+++ b/src/backend/tsearch/wparser_def.c
@@ -7,7 +7,7 @@
  *
  *
  * IDENTIFICATION
- *       $PostgreSQL: pgsql/src/backend/tsearch/wparser_def.c,v 1.31 2010/05/08 16:39:49 tgl Exp $
+ *       $PostgreSQL: pgsql/src/backend/tsearch/wparser_def.c,v 1.32 2010/05/09 02:15:59 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -323,10 +323,9 @@ TParserInit(char *str, int len)
 
 #ifdef WPARSER_TRACE
        /*
-        * Use of %.*s here is not portable when the string contains multibyte
-        * characters: some machines interpret the length in characters, others
-        * in bytes.  Since it's only a debugging aid, we haven't bothered to
-        * fix this.
+        * Use of %.*s here is a bit risky since it can misbehave if the data
+        * is not in what libc thinks is the prevailing encoding.  However,
+        * since this is just a debugging aid, we choose to live with that.
         */
        fprintf(stderr, "parsing \"%.*s\"\n", len, str);
 #endif

diff --git a/src/backend/utils/adt/datetime.c b/src/backend/utils/adt/datetime.c
index 743ca8345d609d96a09bd49cbe674773988dbee9..65d2b875f0794108aabaf96993786c9f88bee65c 100644 (file)
--- a/src/backend/utils/adt/datetime.c
+++ b/src/backend/utils/adt/datetime.c
@@ -8,7 +8,7 @@
  *
  *
  * IDENTIFICATION
- *       $PostgreSQL: pgsql/src/backend/utils/adt/datetime.c,v 1.211 2010/05/08 16:39:51 tgl Exp $
+ *       $PostgreSQL: pgsql/src/backend/utils/adt/datetime.c,v 1.212 2010/05/09 02:15:59 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -3741,11 +3741,9 @@ EncodeDateTime(struct pg_tm * tm, fsec_t fsec, int *tzp, char **tzn, int style,
                        AppendTimestampSeconds(str + strlen(str), tm, fsec);
 
                        /*
-                        * Note: the uses of %.*s in this function would be unportable
-                        * if the timezone names ever contain non-ASCII characters,
-                        * since some platforms think the string length is measured
-                        * in characters not bytes.  However, all TZ abbreviations in
-                        * the Olson database are plain ASCII.
+                        * Note: the uses of %.*s in this function would be risky if the
+                        * timezone names ever contain non-ASCII characters.  However, all
+                        * TZ abbreviations in the Olson database are plain ASCII.
                         */
 
                        if (tzp != NULL && tm->tm_isdst >= 0)

diff --git a/src/bin/psql/print.c b/src/bin/psql/print.c
index 1d73fd5790f39a8c16d6b3e5e94fd3a02ffba86f..86e8a09a13b793b1fcfefed94c12219d07bac5d1 100644 (file)
--- a/src/bin/psql/print.c
+++ b/src/bin/psql/print.c
@@ -3,7 +3,7 @@
  *
  * Copyright (c) 2000-2010, PostgreSQL Global Development Group
  *
- * $PostgreSQL: pgsql/src/bin/psql/print.c,v 1.125 2010/05/08 16:39:52 tgl Exp $
+ * $PostgreSQL: pgsql/src/bin/psql/print.c,v 1.126 2010/05/09 02:15:59 tgl Exp $
  */
 #include "postgres_fe.h"
 
@@ -255,8 +255,8 @@ format_numeric_locale(const char *my_str)
 /*
  * fputnbytes: print exactly N bytes to a file
  *
- * Think not to use fprintf with a %.*s format for this.  Some machines
- * believe %s's precision is measured in characters, others in bytes.
+ * We avoid using %.*s here because it can misbehave if the data
+ * is not valid in what libc thinks is the prevailing encoding.
  */
 static void
 fputnbytes(FILE *f, const char *str, size_t n)

diff --git a/src/interfaces/ecpg/pgtypeslib/dt_common.c b/src/interfaces/ecpg/pgtypeslib/dt_common.c
index dc5fbe7fd96b2b19850f429de6d866e2bc643fe7..52bb5d2ca170a92d5d37f504711de621b06bd1db 100644 (file)
--- a/src/interfaces/ecpg/pgtypeslib/dt_common.c
+++ b/src/interfaces/ecpg/pgtypeslib/dt_common.c
@@ -1,4 +1,4 @@
-/* $PostgreSQL: pgsql/src/interfaces/ecpg/pgtypeslib/dt_common.c,v 1.52 2010/05/08 16:39:52 tgl Exp $ */
+/* $PostgreSQL: pgsql/src/interfaces/ecpg/pgtypeslib/dt_common.c,v 1.53 2010/05/09 02:15:59 tgl Exp $ */
 
 #include "postgres_fe.h"
 
@@ -856,11 +856,9 @@ EncodeDateTime(struct tm * tm, fsec_t fsec, int *tzp, char **tzn, int style, cha
                                sprintf(str + strlen(str), " BC");
 
                        /*
-                        * Note: the uses of %.*s in this function would be unportable
-                        * if the timezone names ever contain non-ASCII characters,
-                        * since some platforms think the string length is measured
-                        * in characters not bytes.  However, all TZ abbreviations in
-                        * the Olson database are plain ASCII.
+                        * Note: the uses of %.*s in this function would be risky if the
+                        * timezone names ever contain non-ASCII characters.  However, all
+                        * TZ abbreviations in the Olson database are plain ASCII.
                         */
 
                        if (tzp != NULL && tm->tm_isdst >= 0)

diff --git a/src/interfaces/libpq/fe-misc.c b/src/interfaces/libpq/fe-misc.c
index 096e0d84a7a1e016751b97c812b90342c0d19694..f76168ccd44cdd493dd0ddf2eaf55f375bf94090 100644 (file)
--- a/src/interfaces/libpq/fe-misc.c
+++ b/src/interfaces/libpq/fe-misc.c
@@ -23,7 +23,7 @@
  * Portions Copyright (c) 1994, Regents of the University of California
  *
  * IDENTIFICATION
- *       $PostgreSQL: pgsql/src/interfaces/libpq/fe-misc.c,v 1.142 2010/05/08 16:39:53 tgl Exp $
+ *       $PostgreSQL: pgsql/src/interfaces/libpq/fe-misc.c,v 1.143 2010/05/09 02:16:00 tgl Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -70,8 +70,8 @@ static int    pqSocketPoll(int sock, int forRead, int forWrite, time_t end_time);
 /*
  * fputnbytes: print exactly N bytes to a file
  *
- * Think not to use fprintf with a %.*s format for this.  Some machines
- * believe %s's precision is measured in characters, others in bytes.
+ * We avoid using %.*s here because it can misbehave if the data
+ * is not valid in what libc thinks is the prevailing encoding.
  */
 static void
 fputnbytes(FILE *f, const char *str, size_t n)