`csv` module should use `PyUnicodeWriter`

[maurycy]

Bug report

Bug description:

csv.writer does a lot of manual buffer manipulations:

• cpython/Modules/_csv.c

  Line 1115 in f58a7c7

  join_reset(WriterObj *self)

• cpython/Modules/_csv.c

  Line 1127 in f58a7c7

  join_append_data(WriterObj *self, int field_kind, const void *field_data,

• cpython/Modules/_csv.c

  Line 1221 in f58a7c7

  join_check_rec_size(WriterObj *self, Py_ssize_t rec_len)

• cpython/Modules/_csv.c

  Line 1240 in f58a7c7

  join_append(WriterObj *self, PyObject *field, int quoted)

• cpython/Modules/_csv.c

  Line 1283 in f58a7c7

  join_append_lineterminator(WriterObj *self)

I believe that big parts of this is no longer needed, since 3.14 and introduction of PyUnicodeWriter:

• https://docs.python.org/dev/c-api/unicode.html#c.PyUnicodeWriter

gh-125196 seems to be a catch-all for using the new API but I believe it's missing the csv module.

CPython versions tested on:

CPython main branch

Operating systems tested on:

Linux

Linked PRs

• gh-138270: Use PyUnicodeWriter in csv.writer #138271

[maurycy]

Is this an issue or a prescription of the implementation?