In article <ALEX.94Nov7145058 at dapsun.lif.icnet.uk>, alex at dapsun.lif.icnet.uk (Alex Whittaker) writes:
> We are offering the GCG 8.0 release on our Digital AXP (Alpha) system
> running OSF/1 and are experiencing some difficulties with the Seqed/lineup
> software run over a Mac Telnet version 2.5 terminal. The problem is
> relates to the arrow keys which the mapping for which is changed from ^[[A
> ^[[B ^[[C and ^[[D to ^[OA ^[OB ^[OC and ^[OD. This is not interpreted
> correctly, and causes the editor to discard the escape O and print a
> capital A B C or D to the screen. As seqed is fairly essential to the
> running of a reasonable service, and most users connect through telnet 2.5
> we are very concerned. I voiced these concerns to GCG and had the
> following response:
>>>>One thing that you can try is:
>>>>Edit Config.tel (on the MAC)
>>>>Then log on to the unix system. For sun set term=vt200 for
>>irix set term=vt220. I'm not sure which to try on OSF. This does not work on
>>Ultrix, and since OSF is also from DEC it may not work their either. (but
>>you can take the sun terminfo entry for vt200 and compile it on the ultrix
>>machine and the arrow keys will work)
>>>>We never tracked down the exact source of the problem, and no longer have
>>version 2.5 for testing.
>>>> Although I appreciate that it is difficult to cater for differences in
> other peoples software, upgrading telnet accross arround 400 macs is not a
> short term solution. The alternative suggested modification to the
> config.tel file has the effect of turning off the arrow keys completely.
>> Has anyone else been able to resolve this issue, could they share it with
> us. If not, can anyone suggest an alternative sequence editor in the
> public domain, compatible with our system.
Macs at our site work happily with NCSA 2.5 and
% setenv TERM ansi
Reason being, that the 'termcap' entry on OSF/1 defines lots of
padding and other things, e.g. a 5 milliseconds padding for relative
cursor screen movement (cm=5\E[%i%d;%dH). The vt200 features are different
and will kill the seqed program also. The 'ansi' definition does not have
any of these features but is a 'ansi terminal with pessimistic assumptions'
- however, I am not aware of possible ramifications of this setting.
Maybe this helps