A list and tuple accept a negative index, slicing doesn't get `index out of range` error and for slicing, the left index can be greater than the right index e.g. `[10:-10]`

[hyperkai]

Bug description:

A list and tuple have some problems as shown below:

*Problems:

1. A negative index is possible to use so it should be impossible to use. *I've never used it in my life.
2. Slicing doesn't get index out of range error while indexing gets it so slicing should also get index out of range error.
3. For slicing, the left index can be greater than the right index, e.g. [10:-10] so they shouldn't be like that.

<List indexing>:

v = ['A', 'B', 'C', 'D', 'E']

print(v[2], v[-3])
# C C

print(v[-10], v[10])
# IndexError: list index out of range

<Tuple indexing>:

v = ('A', 'B', 'C', 'D', 'E')

print(v[2], v[-3])
# C C

print(v[-10], v[10])
# IndexError: tuple index out of range

<List slicing>:

v = ['A', 'B', 'C', 'D', 'E']

print(v[2], v[-3])
# C C

print(v[-10:10]) # No error
# ['A', 'B', 'C', 'D', 'E']

print(v[10:-10]) # No error
# []

<Tuple slicing>:

v = ('A', 'B', 'C', 'D', 'E')

print(v[2], v[-3])
# C C

print(v[-10:10]) # No error
# ('A', 'B', 'C', 'D', 'E')

print(v[10:-10]) # No error
# []

CPython versions tested on:

3.12

Operating systems tested on:

Windows

[gvanrossum]

This behavior has been in the language over 30 years. Every detail of the design exists for a reason. This won't change.

[StanFromIreland]

Please also remember Adam’s suggestions on the tone of your issues.

[tim-one]

The issue tracker isn't the place for tutorials, so if you want more help on this, please consider bringing it up in the "Python Help" topic on discuss.python.org.

Short course: it's been this way "forever", is intentional, won't change, and applies to all sequence types. Lists and tuples, yes, but also strings, bytes, bytearrays, array.arrays. collection.deques, ... Negative indices are in fact heavily used in some code, most typically using sequence[-1] to retrieve "the last" element of sequence.

[skirpichev]

A negative index is possible to use so it should be impossible to use. *I've never used it in my life.

Who are you to teach other people what they should use in their life?!

Slicing doesn't get index out of range error while indexing gets it

Maybe because slicing is for something different, right?

Apparently, you are assumed, that slice arguments are related to a given sequence. But slice is a different object, you can use it (same object) in a subscript notation for various sequences, of different length. It has nothing common with boundaries of the given sequence. Consider this:

>>> v = ['A', 'B', 'C', 'D', 'E']
>>> xs = [1, 2, 3]
>>> myslice = slice(1, None)  # (1)
>>> v[myslice]
['B', 'C', 'D', 'E']
>>> xs[myslice]
[2, 3]
>>> myslice2 = slice(1, 10)  # (2)
>>> v[myslice]
['B', 'C', 'D', 'E']
>>> xs[myslice]
[2, 3]

Of course, all this is documented in depth.

I'm not asking you why (2) "should be" a wrong error (IndexError). But why it should be error at all? slice(1, None) is a way to tell "I want a sub-sequence, starting from index=1". slice(1, 10) is a way to tell, that such sub-sequence always should have upper boundary for it's length.

@hyperkai, please don't open bug reports, until you carefully study documentation. Please ensure, that behavior you treat as "bug" is not documented.

[gvanrossum]

I don’t know what else the OP has filed, but I would write this off to inexperience with Python and a limited grasp of English. Let’s give folks the benefit of the doubt. I f there’s stuff I ought to know DM me on Discord.

[pochmann3]

slice(1) is a way to tell "I want a sub-sequence, starting from index=1"

@skirpichev Did you break your Python somehow? It should be stopping at 1. But I see in your example session that you indeed got starting behavior.

[skirpichev]

Indeed, I copy-pasted wrong input line from REPL (but correct output). Now stop=None - restored. @pochmann3, thanks!