Re: "mini" pascal by Wirth

adrian@platon.cs.rhbnc.ac.uk (A Johnstone) writes:


>BTW an interesting problem we had was that at first the [Pascal-S] parser
>wouldn't work. Like most of Wirth's compilers the keywords are kept in
>an array in alphabetical order and then each alphanumeic ID is checked
>using a binary search on the table. Unfortunately on the machine that
>Pascal-S was developed for DOWNTO came before DO in the lexicographic
>ordering


Yes. It was developed on a Control Data Cyber, which used a six bit
character set.


The alphabetics were 1 through 32. Zero was 33. Blank was 55. (All
these in good old-fashioned octal, of course)


I believe Pascal-S had no pointers. It is twenty years since I looked
at it, so I could be mis-remembering. However, it was written in
standard Pascal, in which allocation of dynamic sized objects is
hard, so I expect my memory is correct on this.


Pascal-S was a compile-and-go program. It compiled into an internal form
and then executed the program. This used to be quite useful on Cybers
which were disk bound when time sharing.
--

Return to the comp.compilers page.
Search the comp.compilers archives again.