On Wed, Jul 24, 2019 at 01:43:23PM -0500, Eric Blake wrote:
On 7/24/19 11:54 AM, Richard W.M. Jones wrote:
> This is a bitmask so using an unsigned type is slightly safer. This
> is not an ABI change since the types are compatible.
>
> Thanks: Eric Blake.
> ---
> examples/strict-structured-reads.c | 2 +-
> generator/generator | 67 ++++++++++++++++--------------
> interop/structured-read.c | 2 +-
> lib/internal.h | 2 +-
> tests/oldstyle.c | 2 +-
> 5 files changed, 39 insertions(+), 36 deletions(-)
>
> @@ -3789,8 +3789,8 @@ let print_python_binding name { args; ret } =
> pr " for (size_t i = 0; i < %s; ++i)\n" len;
> pr " PyList_SET_ITEM (py_%s, i, PyLong_FromUnsignedLong
(%s[i]));\n" n n
> | BytesIn _
> - | Int _
> - | Int64 _ -> ()
> + | Int _
> + | Int64 _ -> ()
A bit of churn here that could have been squashed into patch 2. But at
the end of the day, whether to spend more time polishing the patch when
the end result is fine is up to you.
ACK series. I'll rebase my auto-retire patches on top of this once
you've pushed.
I think I need a bit more time on this series. How about you
push yours and I'll rebase this one on top of it.
Rich.
--
Richard Jones, Virtualization Group, Red Hat
http://people.redhat.com/~rjones
Read my programming and virtualization blog:
http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.
http://fedoraproject.org/wiki/MinGW