Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!newstop!grapevine!mlsgcm!gcmcnutt
From: gcmcnutt@mlsgcm.Ebay.sun.COM (Greg McNutt)
Newsgroups: comp.windows.x
Subject: Is you xrn6.13 crashing too?
Message-ID: <477@grapevine.EBay.Sun.COM>
Date: 23 Oct 90 21:38:10 GMT
Sender: news@grapevine.EBay.Sun.COM
Reply-To: gcmcnutt@mlsgcm.Ebay.sun.COM (Greg McNutt)
Organization: Sun Microsystems Federal, Inc.
Lines: 33

We have been fighting a problem with the latest versions of
xrn and libXt.  It appears that our upgrade from xrn6.10 to
xrn6.13 (or even xrn6.11) now core dumps at an interesting
place.

We are using MIT's libXt for this test (patched to fix-18).
By the way, xrn does not crash with openwindows 2.0's libXt.
This might be because that version of Xt has not been
'patched'...

The crash appears on the *third* entry into the article read
mode.  We have traced back the failure to the *third* entry
back into the newgroup mode i.e.:

run xrn (ng mode) -> read -> ng -> read -> ng

At this point the translations table for the read mode is
XtFree'd during the translations install for the ng mode.

I know, I know, too much detail but after perusing the various
fix-es, we have discovered a rather large amount of 'changes'
to the translation code in *both* xrn and Xt....

We are right now at a loss as to where the problem is and
where the symptoms are.  We suspect that this is a case of
unsynchronized enhancements conflicting with each other.

Any ideas?

greg

BTW: We are using the resources file provided with xrn6.13 for
the 6.13 test....