Repository navigation
Expand file tree
/
Copy path.gitattributes
More file actions
72 lines (72 loc) · 4.44 KB
/
Copy path.gitattributes
File metadata and controls
72 lines (72 loc) · 4.44 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
# Signed-envelope fixtures are byte-exact inputs to signature verification:
# CRLF checkout conversion on Windows changes the signed bytes and breaks
# the generator-match and signature tests.
crates/aft/tests/fixtures/gh_shim/** -text
# Plexus's gh.route contract vectors are parsed byte-for-byte by the shim
# tests; they must stay LF on every platform (Windows autocrlf would
# otherwise check them out CRLF). Later, more specific lines win over the
# -text line above.
crates/aft/tests/fixtures/gh_shim/plexus-contract/** text eol=lf
# ALF's diagnostics.findings vectors are checked against sha256 digests
# recorded from prefrontal's LF bytes, so they must stay LF on Windows too.
crates/aft/tests/fixtures/diagnostics-findings-v1/** text eol=lf
# Binding-discovery vectors and their provenance hashes also pin LF bytes.
crates/aft/tests/fixtures/gh_shim/plexus-bindings-contract/** text eol=lf
crates/aft/tests/fixtures/pty_render/*.raw binary
# Golden render outputs are compared byte-exact against `\n`-joined screen
# output, so they must stay LF on every platform (Windows autocrlf would
# otherwise check them out CRLF and break the deterministic comparison).
crates/aft/tests/fixtures/pty_render/*.xterm.txt eol=lf
# Truncation-contract goldens are compared byte-exact against rendered
# text (`\n` line ends) on every platform; the same Windows autocrlf rule.
crates/aft/tests/fixtures/callgraph/truncation/*.txt text eol=lf
# The tool-result cap golden is compared byte-exact against rendered text.
crates/aft/tests/fixtures/tool-result-cap.txt text eol=lf
# Outline source fixtures and byte-exact renderer goldens use LF even when
# Windows checks out the surrounding repository with core.autocrlf=true.
crates/aft/tests/fixtures/outline_summaries/** text eol=lf
# Bash-surface fixtures: input.txt is the compressor input and is compared
# line-for-line against the rendered output; keep LF on every platform.
crates/aft/tests/fixtures/bash/** text eol=lf
# Inspect renderer goldens are byte-exact pretty JSON with LF line endings.
crates/aft/tests/fixtures/inspect/noise-renderer.json text eol=lf
# The real-query vector pack and the digest-pinned JSON beside it are hashed
# byte-for-byte against pins captured from LF bytes (`embedding_pack_sha256`
# in the manifest, `manifest.sha256`, and the sibling definition digests
# imports_resolve.py checks). A default Windows checkout with
# `core.autocrlf=true` writes CRLF and fails those pins before the gate can
# run, so the checked-out bytes have to be the pinned bytes on every autocrlf
# setting.
benchmarks/aft-search/*.json text eol=lf
# The real-query vector pack is float16 data read byte-for-byte by its digest.
benchmarks/aft-search/*.bin binary
benchmarks/aft-search/manifest.sha256 text eol=lf
# Published exec-remote run-report vectors are pinned by SHA-256 of their
# exact bytes; line endings must never be converted.
crates/aft/src/exec_remote/fixtures/reports/** -text
# Rich-UI canonicalization vectors and templates are digested byte-for-byte:
# the published manifest's digest is the SHA-256 of the bytes, and the phone
# refuses the manifest when its own digest disagrees. They are read with
# include_bytes!, so a Windows checkout with core.autocrlf=true must not
# rewrite them into CRLF. The whole directory is covered so that template
# files added beside the vectors are byte-pinned without another rule.
crates/aft/tests/fixtures/rich_ui/** -text
# Published agent-data project/session vectors are pinned by SHA-256 of their
# exact bytes; Windows checkout conversion must not rewrite their line endings.
crates/aft/tests/fixtures/agent_data/** -text
# Secret-approval and Cingulate secret_use vectors are pinned by the owners'
# sha256 sums of their exact published bytes; Windows autocrlf would rewrite
# them as CRLF on checkout and break the pins.
crates/aft/tests/fixtures/secret/** -text
# Store-diet parity compares raw golden bytes and hashes source-derived facts.
# Preserve fixture/oracle blobs verbatim even with Windows core.autocrlf=true;
# gzip captures must remain binary, and source byte offsets must not drift.
crates/aft/tests/store_diet/*.json -text
crates/aft/tests/store_diet/*.json.gz -text
crates/aft/tests/store_diet/*.patch -text
crates/aft/tests/store_diet/corpus/** -text
crates/aft/tests/store_diet/wire-*/** -text
crates/aft/tests/fixtures/callgraph/** -text
# These full-file inputs are also copied into the frozen language corpus.
crates/aft/tests/fixtures/sample.* -text
crates/aft/tests/fixtures/imports_ts.ts -text