Utilize `bytearray.resize()` and `os.readinto()` to reduce copies
and match behavior of `_io.FileIO.readall()`.
There is still an extra copy which means twice the memory required
compared to FileIO because there isn't a zero-copy path from
`bytearray` -> `bytes` currently.
On my system reading a 2 GB file:
`./python -m test -M8g -uall test_largefile -m test.test_largefile.PyLargeFileTest.test_large_read -v`
Goes from ~2.7 seconds -> ~2.2 seconds
Co-authored-by: Victor Stinner <vstinner@python.org>
Slicing buf and appending chunk would always result in a copy. Commonly
in a readall() there is no already read data in buf, and the amount of
data read may be large, so the copy is expensive.
Both now use a pre-allocated buffer of length `bufsize`, fill it using
a readinto(), and have matching "expand buffer" logic.
On my machine this takes:
`./python -m test -M8g -uall test_largefile -m test_large_read -v`
from ~3.7 seconds to ~3.4 seconds.
`os.read()` allocated and filled a buffer by calling `read(2)`, than that
data was copied into the user provied buffer. Read directly into the
caller's buffer instead by using `os.readinto()`.
`os.readinto()` uses `PyObject_GetBuffer()` to make sure the passed
in buffer is writeable and bytes-like, drop the manual check.
Performed an audit of `fileio.c` and `_pyio` and made sure anytime the
fd changes the stat result, if set, is also cleared/changed.
There's one case where it's not cleared, if code would clear it in
__init__, keep the memory allocated and just do another fstat with the
existing memory.
Spotted by @ngnpope.
`isatty` returns False to indicate the file is not a TTY. The C
implementation of _io does that (`Py_RETURN_FALSE`) but I got the
bool backwards in the _pyio implementaiton.
Multiple places in the I/O stack optimize common cases by using the
information from stat. Currently individual members are extracted from
the stat and stored into the fileio struct. Refactor the code to store
the whole stat struct instead.
Parallels the changes to _io. The `stat` Python object doesn't allow
changing members, so rather than modifying estimated_size, just clear
the value.
This reduces the system call count of a simple program[0] that reads all
the `.rst` files in Doc by over 10% (5706 -> 4734 system calls on my
linux system, 5813 -> 4875 on my macOS)
This reduces the number of `fstat()` calls always and seek calls most
the time. Stat was always called twice, once at open (to error early on
directories), and a second time to get the size of the file to be able
to read the whole file in one read. Now the size is cached with the
first call.
The code keeps an optimization that if the user had previously read a
lot of data, the current position is subtracted from the number of bytes
to read. That is somewhat expensive so only do it on larger files,
otherwise just try and read the extra bytes and resize the PyBytes as
needeed.
I built a little test program to validate the behavior + assumptions
around relative costs and then ran it under `strace` to get a log of the
system calls. Full samples below[1].
After the changes, this is everything in one `filename.read_text()`:
```python3
openat(AT_FDCWD, "cpython/Doc/howto/clinic.rst", O_RDONLY|O_CLOEXEC) = 3`
fstat(3, {st_mode=S_IFREG|0644, st_size=343, ...}) = 0`
ioctl(3, TCGETS, 0x7ffdfac04b40) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
read(3, ":orphan:\n\n.. This page is retain"..., 344) = 343
read(3, "", 1) = 0
close(3) = 0
```
This does make some tradeoffs
1. If the file size changes between open() and readall(), this will
still get all the data but might have more read calls.
2. I experimented with avoiding the stat + cached result for small files
in general, but on my dev workstation at least that tended to reduce
performance compared to using the fstat().
[0]
```python3
from pathlib import Path
nlines = []
for filename in Path("cpython/Doc").glob("**/*.rst"):
nlines.append(len(filename.read_text()))
```
[1]
Before small file:
```
openat(AT_FDCWD, "cpython/Doc/howto/clinic.rst", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=343, ...}) = 0
ioctl(3, TCGETS, 0x7ffe52525930) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
lseek(3, 0, SEEK_CUR) = 0
fstat(3, {st_mode=S_IFREG|0644, st_size=343, ...}) = 0
read(3, ":orphan:\n\n.. This page is retain"..., 344) = 343
read(3, "", 1) = 0
close(3) = 0
```
After small file:
```
openat(AT_FDCWD, "cpython/Doc/howto/clinic.rst", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=343, ...}) = 0
ioctl(3, TCGETS, 0x7ffdfac04b40) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
read(3, ":orphan:\n\n.. This page is retain"..., 344) = 343
read(3, "", 1) = 0
close(3) = 0
```
Before large file:
```
openat(AT_FDCWD, "cpython/Doc/c-api/typeobj.rst", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=133104, ...}) = 0
ioctl(3, TCGETS, 0x7ffe52525930) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
lseek(3, 0, SEEK_CUR) = 0
fstat(3, {st_mode=S_IFREG|0644, st_size=133104, ...}) = 0
read(3, ".. highlight:: c\n\n.. _type-struc"..., 133105) = 133104
read(3, "", 1) = 0
close(3) = 0
```
After large file:
```
openat(AT_FDCWD, "cpython/Doc/c-api/typeobj.rst", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=133104, ...}) = 0
ioctl(3, TCGETS, 0x7ffdfac04b40) = -1 ENOTTY (Inappropriate ioctl for device)
lseek(3, 0, SEEK_CUR) = 0
lseek(3, 0, SEEK_CUR) = 0
read(3, ".. highlight:: c\n\n.. _type-struc"..., 133105) = 133104
read(3, "", 1) = 0
close(3) = 0
```
Co-authored-by: Shantanu <12621235+hauntsaninja@users.noreply.github.com>
Co-authored-by: Erlend E. Aasland <erlend.aasland@protonmail.com>
Co-authored-by: Victor Stinner <vstinner@python.org>
Tools such as ruff can ignore "imported but unused" warnings if a
line ends with "# noqa: F401". It avoids the temptation to remove
an import which is used effectively.
lseek() always returns 0 for character pseudo-devices like
`/dev/urandom` (for other non-regular files, e.g. `/dev/stdin`, it
always returns -1, to which CPython reacts by raising appropriate
exceptions). They are thus technically seekable despite not having seek
semantics.
When calling read() on e.g. an instance of `io.BufferedReader` that
wraps such a file, `BufferedReader` reads ahead, filling its buffer,
creating a discrepancy between the number of bytes read and the internal
`tell()` always returning 0, which previously resulted in e.g.
`BufferedReader.tell()` or `BufferedReader.seek()` being able to return
positions < 0 even though these are supposed to be always >= 0.
Invariably keep the return value non-negative by returning
max(former_return_value, 0) instead, and add some corresponding tests.
Remove io.OpenWrapper and _pyio.OpenWrapper, deprecated in Python
3.10: just use :func:`open` instead. The open() (io.open()) function
is a built-in function. Since Python 3.10, _pyio.open() is also a
static method.
`TextIOWrapper.__init__()` called `os.device_encoding(file.fileno())` if fileno is 0-2 and encoding=None.
But it is very rarely works, and never documented behavior.
open(), io.open(), codecs.open() and fileinput.FileInput no longer
accept "U" ("universal newline") in the file mode. This flag was
deprecated since Python 3.3.
Deprecate io.OpenWrapper and _pyio.OpenWrapper: use io.open and
_pyio.open instead. Until Python 3.9, _pyio.open was not a static
method and builtins.open was set to OpenWrapper to not become a bound
method when set to a class variable. _io.open is a built-in function
whereas _pyio.open is a Python function. In Python 3.10, _pyio.open()
is now a static method, and builtins.open() is now io.open().
The Python _pyio.open() function becomes a static method to behave as
io.open() built-in function: don't become a bound method when stored
as a class variable. It becomes possible since static methods are now
callable in Python 3.10. Moreover, _pyio.OpenWrapper becomes a simple
alias to _pyio.open.
init_set_builtins_open() now sets builtins.open to io.open, rather
than setting it to io.OpenWrapper, since OpenWrapper is now an alias
to open in the io and _pyio modules.
See [PEP 597](https://www.python.org/dev/peps/pep-0597/).
* Add `-X warn_default_encoding` and `PYTHONWARNDEFAULTENCODING`.
* Add EncodingWarning
* Add io.text_encoding()
* open(), TextIOWrapper() emits EncodingWarning when encoding is omitted and warn_default_encoding is enabled.
* _pyio.TextIOWrapper() uses UTF-8 as fallback default encoding used when failed to import locale module. (used during building Python)
* bz2, configparser, gzip, lzma, pathlib, tempfile modules use io.text_encoding().
* What's new entry
The truncate() method of io.BufferedReader() should raise
UnsupportedOperation when it is called on a read-only
io.BufferedReader() instance.
https://bugs.python.org/issue35950
Automerge-Triggered-By: @methane
open(), io.open(), codecs.open() and fileinput.FileInput no longer
accept "U" ("universal newline") in the file mode. This flag was
deprecated since Python 3.3.
* Use the 'p' format unit instead of manually called PyObject_IsTrue().
* Pass boolean value instead 0/1 integers to functions that needs boolean.
* Convert some arguments to boolean only once.
In development mode and in debug build, encoding and errors arguments
are now checked on string encoding and decoding operations. Examples:
open(), str.encode() and bytes.decode().
By default, for best performances, the errors argument is only
checked at the first encoding/decoding error, and the encoding
argument is sometimes ignored for empty strings.
Fix destructor _pyio.BytesIO and _pyio.TextIOWrapper: initialize
their _buffer attribute as soon as possible (in the class body),
because it's used by __del__() which calls close().
Inada Naoki