# Example 2-21. Changing the value of an array item by poking one of its bytes
import array
numbers = array.array('h', [-2, -1, 0, 1, 2])
memv = memoryview(numbers)

Gives me this error:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-3-f1be3db9ba34> in <module>()
----> 1 memv = memoryview(numbers)

TypeError: cannot make memory view because object does not have the buffer interface

Based on research I did, this is a known limitation of python 2.7:

An array can only expose its contents via the old-style buffer interface. This limitation does not apply to Python 3, where memoryview objects can be constructed from arrays, too. Array elements may be multi-byte values.

Source: Python Docs

This info might be useful to someone out there. Not sure if it's worth noting in a future revision of the book :)

First of all - a well written and complete book for intermediate to advanced Python users. It covers a lot of territory and for the most part will work as is on a variety of systems. I did find one bug in chapter 17 that prevents you from moving through the last part of the chapter. The error is in flags2_common.py on Windows 8-10: Not sure if this effects other systems too. The code provided is this...

def main(download_many, default_concur_req, max_concur_req):

    args, cc_list = process_args(default_concur_req)

    actual_req = min(args.max_req, max_concur_req, len(cc_list))

    initial_report(cc_list, actual_req, args.server)

    base_url = SERVERS[args.server]

    t0 = time.time()

    counter = download_many(cc_list, base_url, args.verbose, actual_req)

    assert sum(counter.values()) == len(cc_list), \

        'some downloads are unaccounted for'

    final_report(cc_list, counter, t0)

For windows, this will throw an error: This code works better...

def main(download_many, default_concur_req, max_concur_req):

    args, cc_list = process_args(default_concur_req)

    actual_req = min(args.max_req, max_concur_req, len(cc_list))

    initial_report(cc_list, actual_req, args.server)

    base_url = SERVERS[args.server]

    t0 = time.time()

    counter = download_many(cc_list, base_url, args.verbose, actual_req)

    assert sum(counter.values()) == len(cc_list), 'some downloads are unaccounted for'
    
    final_report(cc_list, counter, t0)

Let me know if you need clarification.

We are no longer inside of the BingoCage class, so we should change the LookupError's wording to reflect that this is an error within the new LotteryBlower class.

Hello sir,

@ramalho I found a GitHub repository with your book downloadable as a .pdf there. I don't know if you allowed that to be put up there, but if not then there's the link: https://github.com/EvanLi/programming-book-3/tree/master/Python

This repository also has dozens of other O'Reilly books for download, not sure if that was intentional but would like to notify you nonetheless. Other than that, I have the paperback book and it is super useful, thank you!

In Python3.7.4

When I test to run this:

>>> hash((1.000001))
2305843009025
>>> hash(1.000001)
2305843009025
>>> (1.000001) == 1.000001
True
>>> [1.000001] == 1.000001
False

But why:

>>> frozenset({1.000001}) == 1.000001
False
>>> hash(frozenset({1.000001}))
8757482414179413028

float, tuple and frozenset are all hashable, why frozenset is special

https://github.com/fluentpython/example-code/blob/3d74f0e18344c2d257e00b689f4c7fe25b974952/06-dp-1class-func/promotions.py#L12

+= should be = if there were 10 item.quantity that bigger than 20 the total would be 0!

While trying to run the tests locally, I ran into errors when pickle tried to load the already existing database. I think it would be nice if readers do not have to waste any time with this error.

This PR fixes that by creating a temporary database before any tests in the module run and deleting it afterwards. I also took the liberty of switching from the deprecated @pytest.yield_fixture to @pytest.fixture.

Thanks for a wonderful book 😃.

the code in listcomp_speed.py will rasise an syntax error: def clock(label, cmd): ... res = time.repeat(cmd, setup=SETUP, number=TIMES) ... print(label, *('{:3.f}.formart(x) for x in res')) File "", line 3 print(label, *('{:3.f}.formart(x) for x in res')) ^ SyntaxError: invalid syntax

but it doesn't make sense to me, as the asterisk in print is unpacking, and unpacking is support in print, as print(*("1","2")) will work well.

and the same code run well in python3.5.

the book did not specify the environment of python.

I was studying docker and used flags example to create and made changes do recreate servers with just one command. @ramalho you can try the instructions on README and check it out. If you like it we I can better reorganize the code ;)

Added support for * specifier in Vector.__format__ to show only small number of components by default, but all if * is specified. Creating this pull request was suggested in the soapbox at the end of Chapter 10.

This request is open for suggestions or improvements and critique!

My VSC runs listcomp_speed.py, it error when running print(label, *('{:.3f}'.format(x) for x in res)) ---- SyntaxError: Generator expression must be parenthesized

Hello, I am a beginner of Python and github. I have just learned chapter 10 of Fluent Python. I noticed that in the last discussion section in chapter 10 you offered a little exercise in adding control over the length of output to the format method. I've just finished this exercise (although I feel like what I added looks jumbled) and added a few simple tests to the doctest (which may not be enough). I hope you can check whether the functions I added are correct. Meanwhile, I hope to take this opportunity to improve my python programming ability and practice the use of git and other functions. Thank you!

I'm trying to run the Example A-2 (page 689 in the First Edition). I'm not sure if it's something wrong that I'm doing.. but I read the code 10 times. I'm sure it is the same of the book.

I'm running this code as:

$python example_a_2__generate_array.py

where my Python version is 3.8.5. Then I see this error:

initial sample: 10500000 elements
Traceback (most recent call last):
  File "example_a_2__generate_array.py", line 19, in <module>
    with len(sample) < SAMPLE_LEN:
AttributeError: __enter__

Some reasoning: I'm not reasigning the open function (as suggested here for one possible cause).

Possible solution: Maybe line 19 should be replaced by:

if len(sample) < SAMPLE_LEN:

I tried that and got a good looking output:

initial sample: 10500000 elements
complete sample: 10500000 elements
not selected: 500000 samples
  writing not_selected.arr
selected: 10000000 samples
  writing selected.arr

And they seem to have been created correctly:

$ls -la *selected*
-rw-rw-r-- 1 lucas lucas  4000000 mai 13 14:57 not_selected.arr
-rw-rw-r-- 1 lucas lucas 80000000 mai 13 14:57 selected.arr

PS.: Thank you for the book and sharing your knowledge!

After that, I can see the rotation in Pycharm. Otherwise, nothing will be seen before "Answer: 42" appear. By the way, the original edition gives what we expected in Jupyter.

my python version is 3.7.6, when I try running tombola_runner.py, it reports to me that: fixture 'cls' not found

the error log is:

file D:\huawei-code\github\example-code\11-iface-abc\tombola_runner.py, line 23
  def test(cls, verbose=False):
E       fixture 'cls' not found
>       available fixtures: cache, capfd, capfdbinary, caplog, capsys, capsysbinary, doctest_namespace, monkeypatch, pytestconfig, record_property, record_testsuite_property, record_xml_attribute, recwarn, tmp_path, tmp_path_factory, tmpdir, tmpdir_factory
>       use 'pytest --fixtures [testpath]' for help on them.