other problems

[talk-2019-ghm-rust.git] / talk.txt

diff --git a/talk.txt b/talk.txt
index 39996abcd824b8f0a0d71d222b268d18970a2dc9..088d44f0c6901b0309bf067a2f98fb51919c2da1 100644 (file)
--- a/talk.txt
+++ b/talk.txt
@@ -250,6 +250,25 @@ are usually based on misunderstandings.)  I guess if your project ends
 up containing piece of opinionated curl-pipe-bash-ware, you should
 expect a culture which produces these kind of problems.
 
+[ other problems ]
+
+Rust does not currently readily support dynamic linking between Rust
+modules.  This is something the Rust community know they need to work
+on but it hasn't been their priority.  Some of Rusts's language design
+choices make this challenging, although not impossible.  In particular
+the very nice system of generics.  Personally I like the tradeoff that
+Rust has made here.  I expect that some form of dynamic linking or
+partial compilation will very likely appear eventually.  But in the
+meantime there are places where replacing C with Rust is awkward.
+
+Rust depends on LLVM.  So its architecture support for smaller
+architectures (with less money behind them) is poor.  From my point of
+view as a Debian contributor a particular problem is the lack of an
+m68k backend for LLVM.  But it provides a more general difficulty with
+supporting a wide set of embedded architectures, which Rust would
+otherwise be very suited to.
+
+
 F======================================================================~
 
 [ rust-lang, ian.jackson@ ]