Moving application from 8.3 to 9.1D (HP-UX)

[jonasj]

Hi guys,

I'm up to moving an application from Progress 8.3 to 9.1D and I'm having trouble.

I'm having code that looks like:

retVal = sqllogin(....); /* retVal == 0 */

EXEC SQL DECLARE myCursor CURSOR FOR

SELECT A, B, C FROM FOO
UNION
SELECT FOO,BAR,BAZ FROM FROTZ
UNION
SELECT GNOME,KDE,CDE FROM DESKTOP
ORDER BY 3,2;
​

and a little later in my pc-file,
EXEC SQL BEGIN DECLARE SECTION;
char i[25];
char j[13];
char k[42];
EXEC SQL END DECLARE SECTION;
EXEC SQL WHENEVER SQLERROR GOTO err;
EXEC SQL OPEN myCursor;
EXEC SQL WHENEVER NOT FOUND GOTO done1;
while(1)
{

EXEC SQL FETCH myCursor INTO :i,:j,:k;
sprintf(outputResult, "%s %s %s", i, j, k);
​

}
done1:
EXEC SQL CLOSE myCursor;
err:
......

This code has been properly compiled and it has been running on HP-UX with Progress 8.3, now, when moving it to Progress 9.1D, and compiling with the provided .o-files and libs, I keep getting coredumps.

gdb reports that it's sqlfetch() that kills me ......

(gdb) ba
#0 0x400000000028c214 in sqlfetch ()
#1 0x40000000000991cc in pat (departm=0x800003ffff581178 "70",
pro_database=0x800003ffff5812f0 "/pathtodatabasefile/db/ade/kurs/pas", resultstr=0x80000001000c63a4 "",
serverlog=0x800003ffff5aa050, threadlog=0x800003ffff5aa070) at pro_mp.c:412
#2 0x40000000000967bc in do_pat (arg=0x80000001000b77f0 "PAT 70 meltest2_syb12\n") at serv_cmds.c:664
#3 0x800003ffff715da0 in __pthread_body () from /usr/lib/pa20_64/libpthread.1
#4 0x800003ffff71f874 in __pthread_start () from /usr/lib/pa20_64/libpthread.1
(gdb) select-frame 1
(gdb) print sqlv
$1 = {0x800003ffff5814b0 "19", 0x800003ffff5814b8 "730120-XXXX", 0x800003ffff5814c8 "Sundgren, Lisa",
0x800003ffff5814f0 "-", 0x800003ffff5814f8 "D", 0x0}
(gdb) print sql0
$2 = 0x80000001000c6370 "0-9289!Sundgren, Lisa", ' ' <repeats 22 times>, "!- !D \n"
(gdb) print SQLCODE
$3 = 429497481496
(gdb) print *SQLCODE
Error accessing memory address 0x64000b7918: Bad address.
(gdb) quit

So is this me, or is there something with this transition that makes this thing crash and die on me all the time???

There seems to come out some data here, is it sqlfetch() that all of a sudden doesn't like to be called repeatedly?
SQLCODE has a very strange value ..... It is declared locally in the function that crash as

long SQLCODE;
​

I hope that somebody can help me with this one .....

 

[jonasj]

SQLCODE messed up ....

Hi guys,

I'm up to moving an application from Progress 8.3 to 9.1D and I'm having trouble.

..............

22 times>, "!- !D \n"
(gdb) print SQLCODE
$3 = 429497481496
(gdb) print *SQLCODE
Error accessing memory address 0x64000b7918: Bad address.
(gdb) quit

So is this me, or is there something with this transition that makes this thing crash and die on me all the time???

There seems to come out some data here, is it sqlfetch() that all of a sudden doesn't like to be called repeatedly?
SQLCODE has a very strange value ..... It is declared locally in the function that crash as

long SQLCODE;
​

I hope that somebody can help me with this one .....

So I helped myself.
It turns out that SQLCODE is zero (0) as long as things work as they should, but instead of becoming something normal, like negative or 100 (sqlcpp translate "EXEC SQL WHENEVER SQLERROR GOTO kloff;" to "if (SQLCODE < 0) goto kloff;" and "EXEC SQL WHENEVER NOT FOUND GOTO koff" turns into "if (sqlcode == 100) goto koff;"), SQLCODE gets a value like 429497481496 or something like that.

Should SQLCODE be declared globally for the whole program? I interpret section 4.1.1 in "Progress Embedded SQL89 Guide and Reference" as if SQLCODE should be declared with local scope.

??