[DBD::Pg] Better logic inside of pg_st_split_statement for the end of the string. Per talk on RT 79035. In a nutshell, we cannot look at *statement, which is one past the current character (ch), if we know what we have already reached the end of the string because ch == '\0'. This was (quite understanably!) causing segfaults on some platforms. Another set of eyes for reviewing welcome (hi Davids)

[dbdpg-commits]

Committed by Greg Sabino Mullane <g...@endpoint.com>

Better logic inside of pg_st_split_statement for the end of the string.
Per talk on RT 79035.
In a nutshell, we cannot look at *statement, which is one past the current
character (ch), if we know what we have already reached the end of the
string because ch == '\0'. This was (quite understanably!) causing segfaults
on some platforms. Another set of eyes for reviewing welcome (hi Davids)

---
 dbdimp.c |    8 ++++++--
 1 files changed, 6 insertions(+), 2 deletions(-)

diff --git a/dbdimp.c b/dbdimp.c
index 84caa8e..ec84e46 100644
--- a/dbdimp.c
+++ b/dbdimp.c
@@ -1973,8 +1973,12 @@ static void pg_st_split_statement (pTHX_ imp_sth_t * 
imp_sth, int version, char
                if (placeholder_type > 0)
                        imp_sth->placeholder_type = placeholder_type;
 
-               /* If this segment also, ended the string, set ch so we bail 
out early */
-               if ('\0' == *statement)
+               /* 
+                  Check if this segment also ends the string.
+                  If it does, we simply leave right away.
+                  Make sure we don't peek at statement if we know it is past 
the end of the string.
+               */
+               if ('\0' != ch && '\0' == *statement)
                        break;
 
        } /* end large while(1) loop: statement parsing */
-- 
1.7.1