diff options
author | Ben Smith <binjimin@gmail.com> | 2017-03-16 14:49:06 -0700 |
---|---|---|
committer | GitHub <noreply@github.com> | 2017-03-16 14:49:06 -0700 |
commit | 5b642c3ce487b77102dec4fc4b55538cfbccc5ff (patch) | |
tree | 960aebda63374bd6d5660355f836e935b402bdc0 /src/binary-reader-objdump.cc | |
parent | 81a64d68c7ce50617bac025061d8f6db91b00ee9 (diff) | |
download | wabt-5b642c3ce487b77102dec4fc4b55538cfbccc5ff.tar.gz wabt-5b642c3ce487b77102dec4fc4b55538cfbccc5ff.tar.bz2 wabt-5b642c3ce487b77102dec4fc4b55538cfbccc5ff.zip |
Use std::unordered_multimap for BindingHash (#357)
This change can't really be done in isolation, since once we add members
with constructors to a struct, it is no longer has a trivial
constructor. This propagates through all types that use it, etc.
There are a number of changes that are ugly, but hopefully reduced the
amount of code that needed to change. In particular, I changed some
union members to pointers so they would stay trivially constructible.
Another tricky change is the handling of duplicate bindings: previously
we just relied on the fact that our hash implementation would be
consistent. A nicer solution is to display the duplicated bindings in
source order. There's probably a nicer way to do this; bikeshedding
welcome. :-)
Diffstat (limited to 'src/binary-reader-objdump.cc')
0 files changed, 0 insertions, 0 deletions