brainstorming around 'simh expect' issue

[Noah Smith]

hi everyone - we've already gotten some great discussion from lars on this topic last friday https://github.com/drforbin/decwar/discussions/26#discussioncomment-13084470

right now we're struggling with some mysterious behaviour - main thing is we want to get our utexas.do simh script 'really robust':) - here's a link to the discussion of this specific topic https://github.com/drforbin/decwar/discussions/28#discussioncomment-13126139 - and a nutshell copypasta to give you the gist of it all right here

----------------------------------------------------------------------------------------------

here are three 'simh .do scripts' - the first two work fine , the third is having kinda 'non deterministic behaviour?' - can we understand why that is?

https://github.com/drforbin/decwar/blob/3eaeaa4760e167cf1fe50bcc9c0298482717c69e/utexas/simh/kermit.do
https://github.com/drforbin/decwar/blob/3eaeaa4760e167cf1fe50bcc9c0298482717c69e/utexas/simh/fort10.do
https://github.com/drforbin/decwar/blob/3eaeaa4760e167cf1fe50bcc9c0298482717c69e/utexas/simh/utexas.do

one interesting note - for kermit, they use the '@script style' for link

expect "\r\n." send "COMP/COMPILE K10UNV.MAC\r"; continue
expect "\r\n." send "COMP/COMPILE K10MIT.MAC,K10SYS.MAC,K10WLD.MAC\r"; continue
expect "\r\n." send "COMP/COMPILE K10MSG.MAC,K10TT.MAC,K10GLB.MAC\r"; continue
expect "\r\n." send "R LINK\r"; continue
expect "\r\n*" send "@K10MIT\r"; continue
expect "\r\n." send "K/F\r"; continue

do remember using this @script style early in the growth of utexas and having basically the same issues as today - but still, it's one thing to investigate.

[Noah Smith]

ahha - one extra note - if you're looking at that utexas.do link, line 11 will probly grab your eye - that's the 'kludge' we've got going to make it work 
and it does - but it shouldn't be necessary!? and we're not confident it works as good on other raspis - major doubts about its 'robusteness'!:)

[Noah Smith]

quick update - we've now got a decent 'fix/workaround/hack' and decwar is now working smoothly 'from zero'

will try to keep following up on root cause, as discussed here https://github.com/drforbin/decwar/discussions/28#discussioncomment-13138478