Re: RE: Question for xt5250

On Fri, 29 Aug 2003, Sean Porterfield wrote:
>
> I couldn't get 0.16.5 to compile on RH9 (I haven't looked into the
> problem yet, I'm likely missing something) but I tried it at home last
> night and got the same result.
>

Frustratingly, RH9 added a requirement to OpenSSL (which the OpenSSL team
told me that they shouldn't have done).  The requirement is to include
the krb5.h header from kerberos -- even if you're not using it -- in order
to make OpenSSL-related programs compile.

That, in itself, wouldn't be a big deal, if the RH9 team had put the
header in the default areas to search for include files!  But, since they
didn't do that, it broke compatibilty with a lot of things.

I wrote a workaround to this problem in CVS, which allows you to make the
CVS versions compile in RH9.   But, 0.16.5 was released before the
workaround was written.

Here's a hack that you can use to make 0.16.5 compile:
http://archive.midrange.com/linux5250/200304/msg00015.html

>
> One thing - a long time ago I had keyboard buffering problems that were
> fixed later (and are still fixed.)  This issue is apparently different
> somehow.  That means I can't tell when it "broke" because it may never
> have worked.  I wasn't using OMS/400 or DBU during that time frame.
>

Do you think it'd be worth trying older versions of TN5250 to see if
there is a version that worked?

If you want to try that, you can do this:

cvs -z3 -d:pserver:anonymous@xxxxxxxxxxxxxxxxxxxxxxxxxx:/cvsroot/tn5250
   co -D 1999-12-05 tn5250

That will check out the version of TN5250 that was in the HEAD branch on
Dec 5th, 1999...

You can also do:

cvs -z3 -d:pserver:anonymous@xxxxxxxxxxxxxxxxxxxxxxxxxx:/cvsroot/tn5250
   co -r v0-16-1 tn5250

to get version 0.16.1 of tn5250.

Using these techniques, you can try older versions and TN5250 and see if
any of them worked.

> I'll see if I can create a program that reproduces the error.

That would probably help.