so the resulting format string is ...%" "PRId64" which is not what we wanted given PRId64 is macro?
The obvious change to use cdata.set doesn't compile without also patching dump.c to #include <inttypes.h>. Not sure why since HAVE_INTTYPES_H is 1.
Using cdata.set_quoted(..., 'lld') compiles without warnings, but ignores all the usual config. That smells like a "works for me" solution :)
While poking around I found a related suggestion in the jansson repo to switch to intmax_t which might also make it easier.
FWIW
Host machine cpu family: x86_64
Host machine cpu: x86_64
jansson| C compiler for the host machine: ccache cc (clang 15.0.0 "Apple clang version 15.0.0 (clang-1500.1.0.2.5)")
Noticed this in passing and it reproduces in this repo. Apologies if I'm missing something obvious...
AFAICT this comes from jansson/meson.build:58
so the resulting format string is
...%" "PRId64"
which is not what we wanted givenPRId64
is macro?The obvious change to use
cdata.set
doesn't compile without also patchingdump.c
to#include <inttypes.h>
. Not sure why sinceHAVE_INTTYPES_H
is 1.Using
cdata.set_quoted(..., 'lld')
compiles without warnings, but ignores all the usual config. That smells like a "works for me" solution :)While poking around I found a related suggestion in the jansson repo to switch to
intmax_t
which might also make it easier.FWIW
Thanks!