Vantage-Config#28's review and issue #1948's comments raised three wording gaps in the
canonical .editorconfig and .gitattributes this repository authors, each surfaced by a
carrier repo that added its own CRLF pin beyond .bat/.cmd:
.editorconfig's "only the CRLF exception below is declared" and its call of
[*.{bat,cmd}] "the one CRLF exception" assume a single exception. A carrier that adds its
own pin (Vantage-Config's .dc) has to reword both locally to stay accurate. Plural phrasing
in the canonical file (e.g. "the CRLF exceptions below") would let such a carrier keep the
comment verbatim..editorconfig's "CI verifies the committed bytes against this file." is inaccurate: CI
runs editorconfig-checker on the checked-out tree, and any text eol=crlf path (the
canonical *.bat/*.cmd, and a carrier's own pin such as Vantage-Config's *.dc) is stored
LF but checked out CRLF. Vantage-Config's local copy already reads "CI verifies the
checked-out files against this file."; a resync should not revert that line until the
canonical wording matches it..gitattributes's header, "Normalize every detected text file to LF in the index and on
checkout.", is inaccurate for the same reason: the *.bat/*.cmd text eol=crlf pins
right below it check those out CRLF. Copilot flagged this on Vantage-Config#31, and
Vantage-Config#32 rewords the local copy to "Normalize every detected text file to LF in the
index, and to LF on checkout except where a CRLF pin below applies." (.gitattributes is
intent fidelity in spec/files.json, so a carrier's local wording is already allowed to
diverge, but the canonical text should say the accurate thing so carriers do not each have to
fix it independently.)
Fixing these is a canonical-content wording change in this repo's own .editorconfig and
.gitattributes, not a per-carrier drift note, and #1948 deliberately left the wording open
rather than deciding it here.
Related: #1948.
🤖 Generated with Claude Code
Vantage-Config#28's review and issue #1948's comments raised three wording gaps in the
canonical
.editorconfigand.gitattributesthis repository authors, each surfaced by acarrier repo that added its own CRLF pin beyond
.bat/.cmd:.editorconfig's "only the CRLF exception below is declared" and its call of[*.{bat,cmd}]"the one CRLF exception" assume a single exception. A carrier that adds itsown pin (Vantage-Config's
.dc) has to reword both locally to stay accurate. Plural phrasingin the canonical file (e.g. "the CRLF exceptions below") would let such a carrier keep the
comment verbatim.
.editorconfig's "CI verifies the committed bytes against this file." is inaccurate: CIruns editorconfig-checker on the checked-out tree, and any
text eol=crlfpath (thecanonical
*.bat/*.cmd, and a carrier's own pin such as Vantage-Config's*.dc) is storedLF but checked out CRLF. Vantage-Config's local copy already reads "CI verifies the
checked-out files against this file."; a resync should not revert that line until the
canonical wording matches it.
.gitattributes's header, "Normalize every detected text file to LF in the index and oncheckout.", is inaccurate for the same reason: the
*.bat/*.cmdtext eol=crlfpinsright below it check those out CRLF. Copilot flagged this on Vantage-Config#31, and
Vantage-Config#32 rewords the local copy to "Normalize every detected text file to LF in the
index, and to LF on checkout except where a CRLF pin below applies." (
.gitattributesisintent fidelity in
spec/files.json, so a carrier's local wording is already allowed todiverge, but the canonical text should say the accurate thing so carriers do not each have to
fix it independently.)
Fixing these is a canonical-content wording change in this repo's own
.editorconfigand.gitattributes, not a per-carrier drift note, and #1948 deliberately left the wording openrather than deciding it here.
Related: #1948.
🤖 Generated with Claude Code