Re: [LINUX5250] Whew! Now we really have it...

[Scott Klement <klemscot@xxxxxxxxxxxx>]

On Fri, 8 Mar 2002, James Rich wrote:

> Ok, I fixed some stupid errors that were preventing multi-page PDFs from
> working.  Thanks go to Scott for one of those (buffer overflow - it wasn't
> the variable gdb was reporting that was having a problem.  variable 'text'
> was being written to way beyond it's bounds and that was wrecking the
> other variables.  Dumb, dumb, dumb).

I didn't look at it that indepth...  clearly a variable was hosed, but
I didn't have the time to trace it down to the source.   Glad you found
it!

> Also the problem with job logs being print all squished should be fixed,
> too.  Thanks to Scott again for pointing me to the AHPP command.
>
> Time to test, test, test!  If anyone wants to try printing directly to PDF
> files, give this a shot.  Some things are not ready yet, but it would be
> good to know what is broken that shouldn't be.

Do you know of any issues with font sizes?   I just did a 1 page document
(a job log) and it worked great.   Then I did a 22 page job log, and it
showed up as a larger font, and chopped the right hand edge off of the
page.

It's progress though!  (Yesterday, it just crashed on the 22 page report!)

> Scott, can you commit these changes?  This should make scs2pdf actually
> work for people, much better than what is already there.
>

Okay, done.

When you do the next batch of changes, do you think you could also add an
entry to the ChangeLog file, and send me a recursive diff?  That way, you
can explain what you did -- rather than me just saying "Applied another
patch from James" which doesn't really give much info.

> Note for developers:
> I placed #ifdef DEBUG ... #endif around fprintf statements that dealt with
> SCS codes that I think we handle correctly.  If you want to see all that
> stuff (and a little more) you need to #define DEBUG.

sounds good.   I wonder if we should consider using syslog(3) for these?
Not just in scs2pdf, but also in scs2ascii & friends...   I just know a
time will come when someone has problems in production, and we need to
track down what's going on.  And that would be hard to do when scs2pdf
is connected to a daemon process by a pipe -- unless we used syslog()...

Not important, though... it was just an idle thought.