[GiNaC-list] Making ginsh interactive when the input is not a TTY

[Richard B. Kreckel]

Dear Vitaly,

You're right about this problem. Thanks a lot for the good analysis! 
Also, your patch seems to address it the right way, in principle. And I 
also agree that the performance hit you describe is not significant.

Could you, please, polish your patch a little? In particular, explain 
the role of the magic '*' char. Also how it works at boundaries of 
YY_READ_BUF_SIZE.

Then, please re-send it as an attachment together with a commit message.

All my best,
    -richy.
-- 
Richard B. Kreckel
<https://in.terlu.de/~kreckel/>