|
View previous topic :: View next topic
|
| Author |
Message |
Kurt Deininger
New User
Joined: 13 Jul 2010
Posts: 19
Location: Frankfurt/Germany
|
|
|
|
| Code: |
The program:
EXEC X1N1168.USER.EXEC(WEIRD) - 01.02 Columns 00001 00072
Command ===> Scroll ===> CSR
****** ***************************** Top of Data ******************************
000001 /* REXX WEIRD */
000002 trace r
000003 nop
=COLS> ----+----1----+----2----+----3----+----4----+----5----+----6----+----7--
000004 lib500="'"f5088(5)".JCLLIBA"'"
000005 fred
000006 lib700="'"f5088(7)".JCLLIB1A"'"
000007 fblc500="FBLC='LC500'"
000008 exit
000009 F5088: return "F5088WDS.RJ.LC"arg(1)"00"
****** **************************** Bottom of Data ****************************
When executing:
3 *-* nop
4 *-* lib500="'"f5088(5)".JCLLIBA"'"
fred
lib700="'"f5088(7)".JCLLIB1A"'"
9 *-* F5088:
*-* return "F5088WDS.RJ.LC"arg(1)"00"
>>> "F5088WDS.RJ.LC500"
>>> "'F5088WDS.RJ.LC500.JCLLIBA"
fred
lib700="f5088(7).JCLLIB1A'"
7 *-* fblc500="FBLC='LC500'"
>>> "FBLC='LC500'"
8 *-* exit
***
|
Hi everybody,
The program was written to generate some JCL. During execution,
some weird things happened, so I cut out the offending part and
created this mini-program. For purpose of illustration I inserted
"fred". Now, this program has a syntax error, if you start counting
the quotes on line 4 (at least I think so). It should give the
message "unmatched quote". Instead, it seems to continue reading the
source past the end of line, despite no comma being present, and
recognize end of of line only at the end of line 6, hence it
resumes execution at line 7. I am puzzled. Is there anything I
overlooked, or does REXX somehow malfunction? You can try it easily
on your mainframe.
Cheers. Kurt. |
|
| Back to top |
|
 |
daveporcelan
Active Member
Joined: 01 Dec 2006
Posts: 792
Location: Pennsylvania
|
|
|
|
There is no malfunction.
There is no unmatched quote.
The translator continued to match up the quotes until it did at the end of line 6.
I use HILITE AUTO and get nice colors to help match things up. |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10872
Location: italy
|
|
|
|
on my pc Object rexx and regina rexx report the error
seems worth investigating with IBM |
|
| Back to top |
|
|
Ronald Burr
Active User
Joined: 22 Oct 2009
Posts: 293
Location: U.S.A.
|
|
|
|
Since REXX recognizes either a single quote or a double quote as literal delimiters, the code shown actually parses like this (if we remove the line numbers) (I've taken the liberty to compress a lot of spaces with <...>):
| Code: |
/* REXX WEIRD */
trace r
nop
=COLS> ---1----+----2----+----3----+----4----+----5----+----6----+----7--
lib500="'"
f5088(5)
".JCLLIBA"
'"<...>fred<...>lib700="' note double quoted string within single quoted string
"f5088(7)"
.JCLLIB1A
"'"
fblc500="FBLC='LC500'" note single quoted string within double quoted string
exit
F5088: return "F5088WDS.RJ.LC"arg(1)"00" |
And there are no unbalanced deliiters.
So, the output trace would parse like this (which matches what you've shown):
| Code: |
3 *-* nop
4 *-* lib500="'"
f5088(5)
".JCLLIBA"
'"<...>fred<...>lib700="'
"f5088(7)"
.JCLLIB1A
"'"
9 *-* F5088:
*-* return "F5088WDS.RJ.LC"arg(1)"00"
>>> "F5088WDS.RJ.LC500"
>>> "'F5088WDS.RJ.LC500.JCLLIBA"<...>fred<...>lib700="f5088(7).JCLLIB1A'"
7 *-* fblc500="FBLC='LC500'"
>>> "FBLC='LC500'"
8 *-* exit |
|
|
| Back to top |
|
|
don.leahy
Active Member
Joined: 06 Jul 2010
Posts: 765
Location: Whitby, ON, Canada
|
|
|
|
| If you use the HILITE REXX command the ISPF editor shows you quite clearly where Rexx thinks the literals are. |
|
| Back to top |
|
|
Kurt Deininger
New User
Joined: 13 Jul 2010
Posts: 19
Location: Frankfurt/Germany
|
|
|
|
Thanks everybody,
After using REXX for 21 years, I find out that (on the mainframe) REXX continues reading over the end of line until it finds a closing delimiter.
(Even says so in the book). All my life I used complete strings with a comma (and concatenation if needed) for long literals.
However, PC-REXX versions don't seem to do that.
I use Quercus REXX privately, and it failed with "unmatched quote", just as Enrico reported.
You learn something new every day (especially when you don't know much in the first place).
Cheers. Kurt |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hi Kurt,
| Quote: |
| You learn something new every day |
One of the better reasons to get out of bed each day. . . 
d |
|
| Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007
Posts: 10872
Location: italy
|
|
|
|
anyway under REXX the continuation has /will always be a murky/unclear issue
if You consider that
| Code: |
parse arg a b
if a = b then
say "equal"
else
say "not equal"
exit
|
| Code: |
parse arg a b
if a = b then ,
say "equal"
else ,
say "not equal"
exit
|
both work the same way...
( at least on object rexx, I do not have TSO up and running right now )
for clarity I resorted to always use a comma for overall consistency!
cheers
enrico |
|
| Back to top |
|
|
|
|
