GitLab

This program contains an expression of the form

   let x = snd (_, snd (_, snd (_, x)))

(probably not explicitly, but that's what appears in the heap at
runtime).  Obviously the program should deadlock if it ever enters
this thing, but apparently the test program in #2327 never does.

The GC tries to evaluate the snd closures, and gets confused due to
the loop.  In particular the earlier fix for #1038 was to blame.

When returning an unboxed tuple with a single non-void component, we
now use the same calling convention as for returning a value of the
same type as that component.  This means that the return convention
for IO now doesn't vary depending on the platform, which make some
parts of the RTS simpler, and fixes a problem I was having with making
the FFI work in unregisterised GHCi (the byte-code compiler makes
some assumptions about calling conventions to keep things simple).

Now ":load M" always searches for a module called "M", rather than
using a file called "M.hs" if that exists.  To get the file semantics
(i.e. not loading "M.o"), use ":load M.hs".

no idea why this only just showed up...

Saves space in the symbol table and speeds up linking

We want to be able to pick the RTS flavour (e.g. -threaded) when we
link the final program.

RTLD_NOW apparently causes some problems, according to previous
mailing-list discussion

 http://www.haskell.org/pipermail/cvs-ghc/2007-September/038570.html

It's now easier to add new repos anywhere in the source tree

just saves a bit of typing

rather than the first one that it finds on the PATH