I recently sent off some nine-track tapes to a fellow named Bob Chicko who had a tape reader and could recover some of the data. These tapes are old and contain some archives from my college days.
I found this email hilarious in its obscurity:Date: 21 Dec 82 15:13:31 PST (Tue)
To: <redacted>
Subject: version of symps
From: Stephen Willson <willson@uci>
Let's see. Right now, there is a generation count stored in the symbol table
descriptor that must match the generation number in the symp for that symp to
be the right one, and not one from another dynamic invocation of a procedure
or function. This was done because the scopep's went away, and we used to
use the scopep's to determine whether a symp was current or not. In explaining
this to Meehan today, it came up, why don't we just store the address of the
current top of the dynamic stack in the symp, and see if these things match.
You see, instead of putting the address of the *scopep* in there, we could just
put the address of the top of the stack (rather than using generation counts).
To: <redacted>
Subject: version of symps
From: Stephen Willson <willson@uci>
Let's see. Right now, there is a generation count stored in the symbol table
descriptor that must match the generation number in the symp for that symp to
be the right one, and not one from another dynamic invocation of a procedure
or function. This was done because the scopep's went away, and we used to
use the scopep's to determine whether a symp was current or not. In explaining
this to Meehan today, it came up, why don't we just store the address of the
current top of the dynamic stack in the symp, and see if these things match.
You see, instead of putting the address of the *scopep* in there, we could just
put the address of the top of the stack (rather than using generation counts).
-- Steve
(It's been a long time since I let anyone call me "Steve".)
No comments:
Post a Comment