iraf.net

Doug (Mink), I just got back from vacation and saw your mail message from last
week regarding lost characters in cursor mode in an Xterm window. If you
have not already figured out the problem based on Suzanne's suggestions,
here are some additional observations. Is this your own version of Xterm,
or the one we distributed under the LOCAL directory with Ultrix/IRAF? The
xterm graphcap entry was designed to go with the latter, although it was
never tested under heavy loading. You may want to modify the graphcap entry
to error-check the returned characters. There is a syntax for pattern matching
in the cursor delimiter parameter. See the comments in gio$stdgraph/stgrcur.x,
and fmtio$patmatch.x. For example, the cursor delimiter and len_curval
parameters (CD and CN respectively) for the vt240 entry are as follows: :CN#-6:\
:CD=?\[[0-9]*,[0-9]*\]^M:CN is negative and the contents of CD are a pattern, in this case because
ReGIS can return different numbers of characters for different coordinates
(but it must return a minimum of 6 characters to be valid). In the example,
the ReGIS cursor return might look like: a[123,456] or
a[1,1]"?" matches the key 'a', "\[" matches the literal '[' (escaped because
'[' is a metacharacter). "[0-9]*" matches any number of digits, "," the ',',
etc., followed by a trailing CR.If characters are being lost, it may be possible to construct a pattern
(with CN negative) that reads up to -CN characters, ignoring the read if the
pattern is not matched, thus avoiding hangup. The user would see that there
was no response and try again.Steve Rooke