byte_array_grow() assigned the result of realloc() directly to
ba->data and increased the capacity even if realloc() failed, so the
following memcpy() in byte_array_append() wrote to NULL + len.

On failure, free the old buffer and leave the byte array in a failed
state (data == NULL), which byte_array_append() already ignores. The
callers check for that state:

- the writers (bin, xml, json, openstep and the text output formats)
  return PLIST_ERR_NO_MEM instead of handing out a NULL/truncated buffer
- node_to_xml() base64-encodes <data> directly into the grown buffer,
  so it must bail out there (only guarded by assert() before)
- the OpenStep parser returns PLIST_ERR_NO_MEM for <hex data> instead of
  silently returning truncated data

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

ptr_array_insert() assigned the result of realloc() directly to
pa->pdata and bumped the capacity without checking for failure, so on
out-of-memory the old buffer was leaked and the following store or
memmove() wrote through a NULL pointer.

Keep the old buffer on failure and return -1 from ptr_array_insert()
and ptr_array_add(). The callers now handle the error:

- the array lookup cache (plist.c) is dropped instead of silently going
  out of sync with the node list; lookups then fall back to walking
  the children, as they do before the cache exists
- bplist parsing (used_indexes) and serialization (objects) return
  PLIST_ERR_NO_MEM; ignoring the error in parse_bin_node_at_index()
  would otherwise loop forever since the array never grows

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

Credit to OSSFuzz

Credit to @ylwango613

Fixes #285

Credit to @ylwango613 for reporting.

- Treat input as unsigned bytes
- Correct UTF-8 bit decoding for 2/3/4-byte sequences
- Add overlong, surrogate, and range checks
- Enforce lead/continuation byte constraints

This addresses issue #283.

Credit to @hgarrereyn for reporting.

Thanks to @unbengable12 for reporting. Addresses #288, #289, #290, #291, and #292.