Re: Memory problems--Carl?

[Harry Binswanger hb@xxxxxxxx]

How large are we
talking about?


I think it's the number of macros, because I'm getting a warning PRompt
message about that, telling me to "clear some macros." The pgm
occupies 12k of disk space and uses about 60 different save/gets. I know
that's too many, so I want a tool that will help me decide which can be
re-used. Would RENUMBER be useful?


 That could be
your problem right

there. If your code is very long, try breaking the routine into

smaller discrete operations -- separate PMs that can be RUN one at
a

time.


Yeah, that's what I have done already--what should be 1 pgm is now 5. But
some of those 5 are getting too long! The length has come because I'm
using these programs to convert plain-text emails into precisely
controlled HTML emails, with tables galore. Took me 7 solid months to
figure out how to make HTML work in all the different, and noncompliant,
email clients. A nightmare. Now I'm just tweaking, almost for
fun.




If the routine really
has to work as an integral PM, issue

KILLMEM, to free up available S/G memory, before you
RUN

the PM.


I already had JM 2. CLEARSGTSQ2 at the start. Just added a call to
KILLMEM, but no improvement. What happens is that s/g 02 gets overwritten
in the HIDE routine (ironic, huh?). After the "clear some
macros" message, I got a message involving s/g 741. So I think what
was supposed to go into 741 couldn't and got put into 02 instead. The
context is:



JM 2.HIDE:01-49,51-89Q2 ;*;

JM 2.DAYSQ2 ;*;

JM 2.UNHIDEQ2 ;*;


I use that to subtract 21 days from the current date. So the content of
50 right now is:

9/22/2011,-21

and that does function correctly, but going into that call, 02 was
"m" and coming out 02 is "4".


I




Hard to say more
without seeing the code. There may be one flawed

operation that's causing the whole thing to tank. Or you may have
to

rethink the design from scratch. I assume you're routinely using

frames IFEI, SYNTAX and RELABEL to check for rudimentary
errors.


Nope, but I will now. Results: all passed, except if I try to do

RELABEL/NR

I get: Bad Path|File




XRAY and XPOLL will
display Save/Get contents while the PM is

RUNning and immediately after it EXits. Viewing the contents of
S/Gs

can help to visualize errors that are hard to catch just by

inspecting the code.



I'd like to use XPOLL, but I'm unclear how to. I tried running the pgm
then doing XPOLL but that didn't show anything (except the
small $X stuff I used instead of literal ). I tried
putting a JM 2.XPOLLQ2 just before the pgm exits, and that produced
nothing.




And believe me, you don't want to see this code.


Thanks for your hand-holding.





Hope this helps.


--

Carl Distefano

cld@xxxxxxxx




Harry Binswanger
hb@xxxxxxxx