Build with clang, and the core and store on macOS, again

Apple clang 21 (an M3) and clang 19 refused what gcc let pass (research
029):
- PURE_FIELD_REST was an enum of UINT32_MAX, past an int (C17): a macro.
- a store's publish copied an entry rebuild left unset when the record
  was not found (the copy was never used): copied when it was made.
- the log's image writer kept a count only Linux's sync_file_range uses.
- ",{" + (row == from) read as adding an int to a string: said plainly.
- the tests: mkdtemp, hidden by _POSIX_C_SOURCE on macOS, needs
  _DARWIN_C_SOURCE; a string split so that \x80 and a stay apart is
  bracketed; a count left over and a malloc measure only glibc has.
The core's and the store's tests pass on the Mac.

1 files changed, 1 insertions, 1 deletions

diff --git a/serve/serve.c b/serve/serve.c
index 81a338d..65a5c0b 100644
--- a/serve/serve.c
+++ b/serve/serve.c
@@ -1429,7 +1429,7 @@ static PureResult show_given(PureApp *app, const Route *route, size_t available,
         char head[80];
         put_bytes(&o, head, (size_t)snprintf(head, sizeof(head), "{\"rows\":%zu,\"from\":%zu,\"items\":[", rows, from));
         for (size_t row = from; row < end && !o.failed; ++row) {
-            put_bytes(&o, ",{" + (row == from), 2 - (row == from));
+            put_bytes(&o, row == from ? "{" : ",{", row == from ? 1 : 2);
             for (unsigned k = 0; k < count; ++k) {
                 put_bytes(&o, columns[k].key + !k, columns[k].length - !k);
                 uint8_t kind = m->view_type[columns[k].view];