`gettext` incorrectly escapes trailing backslashes

[ZeroIntensity]

Describe the bug

Sphinx's gettext extraction escapes trailing backslashes into \\, while the HTML correctly merges them with the next line.

How to Reproduce

$ echo "My hovercraft is \\ \nfull of eels." > index.rst
$ make gettext
$ tail -2 _build/gettext/index.pot
msgid "My hovercraft is \\ full of eels."
msgstr ""

Environment Information

Platform:              linux; (Linux-6.13.6-arch1-1-x86_64-with-glibc2.41)
Python version:        3.13.2 (main, Feb  5 2025, 08:05:21) [GCC 14.2.1 20250128])
Python implementation: CPython
Sphinx version:        8.2.3
Docutils version:      0.21.2
Jinja2 version:        3.1.6
Pygments version:      2.19.1

Sphinx extensions

None

Additional context

Noticed in python/cpython#135781
cc @picnixz

[picnixz]

The change is likely to affect

sphinx/sphinx/util/nodes.py

Line 314 in c384ab9

msg = node.rawsource.replace('\n', ' ').strip() # type: ignore[attr-defined]

where we would first split by lines, then remove trailing backslashes, and rejoin, and then drop lines.

It can however slow down the build a lot as the raw sources may be big... And just replacing all backslashes seem wrong as they could be in the middle of a sentence or escaped =/

[picnixz]

Fixed in #13686.