On Thu, Aug 26, 1999 at 10:50:01PM +0200, Matthias Drochner wrote:
> Afaik it is caused by a delayed TCP ack optimization which came
> between 1.3 and 1.4. A larger vi's "escapetime" helps.
Yes. Thank you Matthias and Erich, a escapetime of 5 works around
this problem.

As Erich has written, it's indeed a split in the ESC sequence that
happens quite often:
17:48:42.507758 localhost.966 > localhost.login: P 88:89(1) ack 594 win 16384 <nop,nop,timestamp 854196 854183> [tos 0x10]
                         4510 0035 2c81 0000 4006 5030 7f00 0001
                         7f00 0001 03c6 0201 a113 5460 a2f1 9bfa
                         8018 4000 d215 0000 0101 080a 000d 08b4
                         000d 08a7 1b
                                   ^^
17:48:42.691806 localhost.966 > localhost.login: P 89:91(2) ack 594 win 16384 <nop,nop,timestamp 854196 854183> [tos 0x10]
                         4510 0036 2c83 0000 4006 502d 7f00 0001
                         7f00 0001 03c6 0201 a113 5461 a2f1 9bfa
                         8018 4000 91d2 0000 0101 080a 000d 08b4
                         000d 08a7 5b41
                                   ^^^^

This "split" is of course completely independent of vi, but I haven't found
another applications so far that has problems with it.

Still strange to me that telnet never does this split...
-- 
Dies ist Thilos Unix Signature! Viel Spass damit.