🚀 go-pugleaf

In article <18p7lpINNc73@almaak.usc.edu> ajayshah@almaak.usc.edu
(Ajay Shah) writes:
>cameron@symcom.math.uiuc.edu (Cameron Smith) [That's me!] writes:
>>Are there enough free Big TeX implementation available nowadays that
>>I can simply tell someone who's running out of memory to go get one?
>
>I find it hard to appreciate why TeX should not know how to dynamically
>allocate RAM on demand.  Does this have something to do with being
>written in pascal?  I seem to recall that new() returns nil if
>you are out of memory.

I don't have the source code to TeX (Volume B of Computers & Typesetting)
available just now (I returned it to the library; couldn't justify buying
a copy), and I didn't study it closely when I did have it, but I seem
to recall DEK's having made a comment that he wasn't happy about
having to statically pre-allocate tables to hold all of TeX's
data structures, but that since Pascal implementations were so
non-standard in their handling of new() he had no choice.  In particular
I believe he specifically stated that on some Pascals a call to new()
when there wasn't enough memory available would either crash the program
or return a pointer to random garbage.  Of course the Pascal you use
today may do something more sensible, but (a) this was about 15 years
ago, and (b) he was writing for the lowest common denominator, trying
to allow for worst-case behavior.  (And a darn good thing too; that's
probably the reason that it has been possible to port TeX to so many
systems having so little in common with one another.)

>Ajay Shah, (213)749-8133, ajayshah@usc.edu

Cameron Smith (cameron@symcom.math.uiuc.edu)