Skip to content

Commit

Permalink
Remove mention of requirements.in from 735
Browse files Browse the repository at this point in the history
This is extending the PEP text for an ancillary point of confusion --
there are tool users for `uv` (and possibly `pip-tools`) who do not
understand that `requirements.in` is a convention that does not
represent a significantly distinct file-format.

Discussing the difference between these only lengthens the already
very-long PEP text, and does not add value vis-a-vis community
standards. Given that neither `requirements.in` nor `requirements.txt`
have clear specifications, other than "what `pip` accepts", keeping
with the older and more standardized naming keeps things shorter and
simpler.
  • Loading branch information
sirosen committed Sep 24, 2024
1 parent 4b98785 commit 608b850
Showing 1 changed file with 0 additions and 15 deletions.
15 changes: 0 additions & 15 deletions peps/pep-0735.rst
Original file line number Diff line number Diff line change
Expand Up @@ -570,21 +570,6 @@ defined dynamically. Requirements loaded from ``requirements.txt`` files and
definitions of static lists prior to ``setup()`` invocation readily analogize
with Dependency Groups.

requirements.txt vs requirements.in
-----------------------------------

Using tools such as ``pip-tools``'s ``pip-compile`` or ``poetry``'s
``poetry export``, many users produce ``requirements.txt`` files which contain
*locked* requirements data, with exact versions and hashes pinned for an entire
environment.

A community standard which is promoted in several projects' documentation,
including that of ``pip-tools``, has evolved of calling the input (unlocked)
requirements data ``requirements.in`` and the locked data ``requirements.txt``.

Users familiar with this usage should typically think of Dependency Groups as
their ``requirements.in`` data, not their locked dependency data.

Interfaces for Use of Dependency Groups
---------------------------------------

Expand Down

0 comments on commit 608b850

Please sign in to comment.