Re: Compiler bootstrapping

drw@euler.mit.edu (Dale R. Worley)
Sun, 29 Aug 1993 02:22:04 GMT

          From comp.compilers
Related articles
Compiler bootstrapping ramana@v01.svl.cdc.com (1993-08-17)
Re: Compiler bootstrapping drw@euler.mit.edu (1993-08-29)
| List of all articles for this month |
Newsgroups: comp.compilers
From: drw@euler.mit.edu (Dale R. Worley)
Keywords: design
Organization: MIT Dept. of Tetrapilotomy, Cambridge, MA, USA
References: 93-08-094
Date: Sun, 29 Aug 1993 02:22:04 GMT

  ramana@v01.svl.cdc.com writes:
[re repetitively feeding a compiler to itself]
        b) Is the process in second approach guaranteed to terminate?
              Under some conditions?


There is a famous case which demonstrates that even if the process
terminates, there is no guarantee that it always terminates in the
same state (i.e., there may be more than one fixed point to the
process): Some famous Unix guy constructed a trojan horse in the C
compiler whose sole function was to insert the trojan horse in the C
compiler when it compiled itself! Once the trojan horse was in the
compiler binary, removing it from the source would not cause it to go
away...


Dale


Dale Worley Dept. of Math., MIT drw@math.mit.edu
[It was Ken Thompson, see ``Reflections on Trusting Trust'' -John]


--

Post a followup to this message

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