[GTALUG] WSL, threat or boon? [was Re:Surveillance Capitalism [was another thread]]
Jamon Camisso
jamon.camisso at utoronto.ca
Mon Apr 5 11:20:35 EDT 2021
On 05/04/2021 09:31, Russell Reiter via talk wrote:
> In that case, in the debate between GCC and clang LLVM, as someone who is
> unable to write an operating system from scratch; who relies on
> documentation
> and the help of like minded people; my vote goes to GCC. It preserves
> support for what
> I see as program necessary artifacts. Plus I see python and other
> interpretative hooks
> into machine code a risk, which must be well balanced, from a SigInt
> perspective.
1. What's the debate about? Links please.
2. What do you mean by interpretative hooks? What is the risk model that
you are conflating with with LLVM, and how is it any different than GCC?
Do you verify all your binaries and compiler and all the intermediate
objects when you build software? As Ken Thompson said, "You can't trust
code that you did not totally create yourself... No amount of
source-level verification or scrutiny will protect you from using
untrusted code[1]."
Since the "debate" as presented here is framed in terms of (specious
until proven otherwise) risk, I suggest that focusing on the compiler is
a secondary concern to the main trust issues that must be addressed,
which are formal verification and reproducible builds. Perhaps the
CompCert compiler would be better for your needs[2].
> In such a case of reconstructionism, I believe GCC is the better
> philosophical option.
Why do you believe it is better? Is using LLVM restricting developers
from writing software that can create social change? Does GCC somehow
better enable developers to engage in critical thinking about the world?
Is any of the above the reason that you use a compiler or write
software? I'd like to understand how either compiler helps or hinders
you, or other developers.
[1] https://dl.acm.org/doi/10.1145/358198.358210
[2] https://github.com/AbsInt/CompCert
More information about the talk
mailing list