jj-vcs
Spurious `jj status`/`jj diff` due to case-insensitivity, such as for gecko-dev
Description
Steps to Reproduce the Problem
- Use Git to clone https://github.com/mozilla/gecko-dev/commit/f0967c33054b3d7c5f9b5e4cf4ae1a85d9ded553
- Initialize colocated repo with
jj init --git-repo . - Run
jj status
Expected Behavior
Fresh clone should be clean.
Actual Behavior
git status is clean:
$ git status
On branch master
Your branch is up to date with 'origin/master'.
nothing to commit, working tree clean
But not jj:
Status
Parent commit: f0967c33054b Backed out changeset 8be4693aecbb (bug 1754905) for causing bustages in xpcAccessiblePivot.h. CLOSED TREE
Working copy : 4354f8335918 (no description set)
Working copy changes:
A testing/web-platform/meta/WebIDL/current-realm.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/allow-resizable.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/builtin-function-properties.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/class-string-interface.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/class-string-iterator-prototype-object.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/class-string-named-properties-object.window.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/constructors.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/default-iterator-object.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/default-toJSON-cross-realm.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/es-exceptions/DOMException-constructor-and-prototype.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/es-exceptions/DOMException-constructor-behavior.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/es-exceptions/DOMException-custom-bindings.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/es-exceptions/exceptions.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/global-immutable-prototype.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/global-object-implicit-this-value-cross-realm.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/global-object-implicit-this-value.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/has-instance.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/interface-object-set-receiver.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/interface-object.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/interface-prototype-constructor-set-receiver.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/interface-prototype-object.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/invalid-this-value-cross-realm.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/iterator-invalidation-foreach.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/iterator-prototype-object.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/legacy-callback-interface-object.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/legacy-factor-function-subclass.window.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/no-regexp-special-casing.any.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/observable-array-no-leak-of-internals.window.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/observable-array-ownkeys.window.js.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/sequence-conversion.html.ini
A testing/web-platform/meta/WebIDL/ecmascript-binding/window-named-properties-object.html.ini
A testing/web-platform/meta/WebIDL/idlharness-shadowrealm.window.js.ini
A testing/web-platform/meta/WebIDL/idlharness.any.js.ini
R testing/web-platform/meta/webidl/current-realm.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/allow-resizable.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/builtin-function-properties.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/class-string-interface.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/class-string-iterator-prototype-object.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/class-string-named-properties-object.window.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/constructors.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/default-iterator-object.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/default-toJSON-cross-realm.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/es-exceptions/DOMException-constructor-and-prototype.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/es-exceptions/DOMException-constructor-behavior.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/es-exceptions/DOMException-custom-bindings.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/es-exceptions/exceptions.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/global-immutable-prototype.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/global-object-implicit-this-value-cross-realm.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/global-object-implicit-this-value.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/has-instance.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/interface-object-set-receiver.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/interface-object.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/interface-prototype-constructor-set-receiver.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/interface-prototype-object.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/invalid-this-value-cross-realm.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/iterator-invalidation-foreach.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/iterator-prototype-object.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/legacy-callback-interface-object.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/legacy-factor-function-subclass.window.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/no-regexp-special-casing.any.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/observable-array-no-leak-of-internals.window.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/observable-array-ownkeys.window.js.ini
R testing/web-platform/meta/webidl/ecmascript-binding/sequence-conversion.html.ini
R testing/web-platform/meta/webidl/ecmascript-binding/window-named-properties-object.html.ini
R testing/web-platform/meta/webidl/idlharness-shadowrealm.window.js.ini
R testing/web-platform/meta/webidl/idlharness.any.js.ini
The diff appears to show removing and adding the same file contents:
Diff
Added regular file testing/web-platform/meta/WebIDL/current-realm.html.ini:
1: [current-realm.html]
2: [getImageData]
3: expected: FAIL
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/allow-resizable.html.ini:
1: [allow-resizable.html]
2: [APIs without [AllowResizable\] throw when passed resizable ArrayBuffers]
3: expected: FAIL
4:
5: [APIs with [AllowShared\] but without [AllowResizable\] throw when passed growable SharedArrayBuffers]
6: expected: FAIL
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/builtin-function-properties.any.js.ini:
1: [builtin-function-properties.any.html]
2:
3: [builtin-function-properties.any.worker.html]
4: expected:
5: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/class-string-interface.any.js.ini:
1: [class-string-interface.any.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [class-string-interface.any.worker.html]
6: expected:
7: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/class-string-iterator-prototype-object.any.js.ini:
1: [class-string-iterator-prototype-object.any.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [class-string-iterator-prototype-object.any.worker.html]
6: expected:
7: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/class-string-named-properties-object.window.js.ini:
1: [class-string-named-properties-object.window.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/constructors.html.ini:
1: [constructors.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/default-iterator-object.html.ini:
1: [default-iterator-object.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/default-toJSON-cross-realm.html.ini:
1: [default-toJSON-cross-realm.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/es-exceptions/DOMException-constructor-and-prototype.any.js.ini:
1: [DOMException-constructor-and-prototype.any.worker.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [DOMException-constructor-and-prototype.any.html]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/es-exceptions/DOMException-constructor-behavior.any.js.ini:
1: [DOMException-constructor-behavior.any.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [DOMException-constructor-behavior.any.worker.html]
6: expected:
7: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/es-exceptions/DOMException-custom-bindings.any.js.ini:
1: [DOMException-custom-bindings.any.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [DOMException-custom-bindings.any.worker.html]
6: expected:
7: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/es-exceptions/exceptions.html.ini:
1: [exceptions.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/global-immutable-prototype.any.js.ini:
1: [global-immutable-prototype.any.worker.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [global-immutable-prototype.any.serviceworker.html]
6: expected:
7: if (os == "android") and fission: [OK, TIMEOUT]
8:
9: [global-immutable-prototype.any.html]
10: expected:
11: if (os == "android") and fission: [OK, TIMEOUT]
12:
13: [global-immutable-prototype.any.sharedworker.html]
14: expected:
15: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/global-object-implicit-this-value-cross-realm.html.ini:
1: [global-object-implicit-this-value-cross-realm.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/global-object-implicit-this-value.any.js.ini:
1: [global-object-implicit-this-value.any.worker.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [global-object-implicit-this-value.any.serviceworker.html]
6: expected:
7: if (os == "android") and fission: [OK, TIMEOUT]
8:
9: [global-object-implicit-this-value.any.html]
10: expected:
11: if (os == "android") and fission: [OK, TIMEOUT]
12:
13: [global-object-implicit-this-value.any.sharedworker.html]
14: expected:
15: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/has-instance.html.ini:
1: [has-instance.html]
2: prefs: [dom.webidl.crosscontext_hasinstance.enabled:false]
3: expected:
4: if (os == "android") and fission: [TIMEOUT, OK]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/interface-object-set-receiver.html.ini:
1: [interface-object-set-receiver.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/interface-object.html.ini:
1: [interface-object.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/interface-prototype-constructor-set-receiver.html.ini:
1: [interface-prototype-constructor-set-receiver.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/interface-prototype-object.html.ini:
1: [interface-prototype-object.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/invalid-this-value-cross-realm.html.ini:
1: [invalid-this-value-cross-realm.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/iterator-invalidation-foreach.html.ini:
1: [iterator-invalidation-foreach.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/iterator-prototype-object.html.ini:
1: [iterator-prototype-object.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/legacy-callback-interface-object.html.ini:
1: [legacy-callback-interface-object.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/legacy-factor-function-subclass.window.js.ini:
1: [legacy-factor-function-subclass.window.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/no-regexp-special-casing.any.js.ini:
1: [no-regexp-special-casing.any.worker.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [no-regexp-special-casing.any.html]
6: expected:
7: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/observable-array-no-leak-of-internals.window.js.ini:
1: [observable-array-no-leak-of-internals.window.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/observable-array-ownkeys.window.js.ini:
1: [observable-array-ownkeys.window.html]
2: prefs: [layout.css.constructable-stylesheets.enabled:true]
3: expected:
4: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/sequence-conversion.html.ini:
1: [sequence-conversion.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
Added regular file testing/web-platform/meta/WebIDL/ecmascript-binding/window-named-properties-object.html.ini:
1: [window-named-properties-object.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4: [[[OwnPropertyKeys\]\]]
5: expected: FAIL
Added regular file testing/web-platform/meta/WebIDL/idlharness-shadowrealm.window.js.ini:
1: [idlharness-shadowrealm.window.html]
2: prefs: [javascript.options.experimental.shadow_realms:true]
3: [DOMException interface: existence and properties of interface object]
4: expected: FAIL
5:
6: [DOMException interface object length]
7: expected: FAIL
8:
9: [DOMException interface object name]
10: expected: FAIL
11:
12: [DOMException interface: existence and properties of interface prototype object]
13: expected: FAIL
14:
15: [DOMException interface: existence and properties of interface prototype object's "constructor" property]
16: expected: FAIL
17:
18: [DOMException interface: existence and properties of interface prototype object's @@unscopables property]
19: expected: FAIL
20:
21: [DOMException interface: attribute name]
22: expected: FAIL
23:
24: [DOMException interface: attribute message]
25: expected: FAIL
26:
27: [DOMException interface: attribute code]
28: expected: FAIL
29:
30: [DOMException interface: constant INDEX_SIZE_ERR on interface object]
31: expected: FAIL
32:
33: [DOMException interface: constant INDEX_SIZE_ERR on interface prototype object]
34: expected: FAIL
35:
36: [DOMException interface: constant DOMSTRING_SIZE_ERR on interface object]
37: expected: FAIL
38:
39: [DOMException interface: constant DOMSTRING_SIZE_ERR on interface prototype object]
40: expected: FAIL
41:
42: [DOMException interface: constant HIERARCHY_REQUEST_ERR on interface object]
43: expected: FAIL
44:
45: [DOMException interface: constant HIERARCHY_REQUEST_ERR on interface prototype object]
46: expected: FAIL
47:
48: [DOMException interface: constant WRONG_DOCUMENT_ERR on interface object]
49: expected: FAIL
50:
51: [DOMException interface: constant WRONG_DOCUMENT_ERR on interface prototype object]
52: expected: FAIL
53:
54: [DOMException interface: constant INVALID_CHARACTER_ERR on interface object]
55: expected: FAIL
56:
57: [DOMException interface: constant INVALID_CHARACTER_ERR on interface prototype object]
58: expected: FAIL
59:
60: [DOMException interface: constant NO_DATA_ALLOWED_ERR on interface object]
61: expected: FAIL
62:
63: [DOMException interface: constant NO_DATA_ALLOWED_ERR on interface prototype object]
64: expected: FAIL
65:
66: [DOMException interface: constant NO_MODIFICATION_ALLOWED_ERR on interface object]
67: expected: FAIL
68:
69: [DOMException interface: constant NO_MODIFICATION_ALLOWED_ERR on interface prototype object]
70: expected: FAIL
71:
72: [DOMException interface: constant NOT_FOUND_ERR on interface object]
73: expected: FAIL
74:
75: [DOMException interface: constant NOT_FOUND_ERR on interface prototype object]
76: expected: FAIL
77:
78: [DOMException interface: constant NOT_SUPPORTED_ERR on interface object]
79: expected: FAIL
80:
81: [DOMException interface: constant NOT_SUPPORTED_ERR on interface prototype object]
82: expected: FAIL
83:
84: [DOMException interface: constant INUSE_ATTRIBUTE_ERR on interface object]
85: expected: FAIL
86:
87: [DOMException interface: constant INUSE_ATTRIBUTE_ERR on interface prototype object]
88: expected: FAIL
89:
90: [DOMException interface: constant INVALID_STATE_ERR on interface object]
91: expected: FAIL
92:
93: [DOMException interface: constant INVALID_STATE_ERR on interface prototype object]
94: expected: FAIL
95:
96: [DOMException interface: constant SYNTAX_ERR on interface object]
97: expected: FAIL
98:
99: [DOMException interface: constant SYNTAX_ERR on interface prototype object]
100: expected: FAIL
101:
102: [DOMException interface: constant INVALID_MODIFICATION_ERR on interface object]
103: expected: FAIL
104:
105: [DOMException interface: constant INVALID_MODIFICATION_ERR on interface prototype object]
106: expected: FAIL
107:
108: [DOMException interface: constant NAMESPACE_ERR on interface object]
109: expected: FAIL
110:
111: [DOMException interface: constant NAMESPACE_ERR on interface prototype object]
112: expected: FAIL
113:
114: [DOMException interface: constant INVALID_ACCESS_ERR on interface object]
115: expected: FAIL
116:
117: [DOMException interface: constant INVALID_ACCESS_ERR on interface prototype object]
118: expected: FAIL
119:
120: [DOMException interface: constant VALIDATION_ERR on interface object]
121: expected: FAIL
122:
123: [DOMException interface: constant VALIDATION_ERR on interface prototype object]
124: expected: FAIL
125:
126: [DOMException interface: constant TYPE_MISMATCH_ERR on interface object]
127: expected: FAIL
128:
129: [DOMException interface: constant TYPE_MISMATCH_ERR on interface prototype object]
130: expected: FAIL
131:
132: [DOMException interface: constant SECURITY_ERR on interface object]
133: expected: FAIL
134:
135: [DOMException interface: constant SECURITY_ERR on interface prototype object]
136: expected: FAIL
137:
138: [DOMException interface: constant NETWORK_ERR on interface object]
139: expected: FAIL
140:
141: [DOMException interface: constant NETWORK_ERR on interface prototype object]
142: expected: FAIL
143:
144: [DOMException interface: constant ABORT_ERR on interface object]
145: expected: FAIL
146:
147: [DOMException interface: constant ABORT_ERR on interface prototype object]
148: expected: FAIL
149:
150: [DOMException interface: constant URL_MISMATCH_ERR on interface object]
151: expected: FAIL
152:
153: [DOMException interface: constant URL_MISMATCH_ERR on interface prototype object]
154: expected: FAIL
155:
156: [DOMException interface: constant QUOTA_EXCEEDED_ERR on interface object]
157: expected: FAIL
158:
159: [DOMException interface: constant QUOTA_EXCEEDED_ERR on interface prototype object]
160: expected: FAIL
161:
162: [DOMException interface: constant TIMEOUT_ERR on interface object]
163: expected: FAIL
164:
165: [DOMException interface: constant TIMEOUT_ERR on interface prototype object]
166: expected: FAIL
167:
168: [DOMException interface: constant INVALID_NODE_TYPE_ERR on interface object]
169: expected: FAIL
170:
171: [DOMException interface: constant INVALID_NODE_TYPE_ERR on interface prototype object]
172: expected: FAIL
173:
174: [DOMException interface: constant DATA_CLONE_ERR on interface object]
175: expected: FAIL
176:
177: [DOMException interface: constant DATA_CLONE_ERR on interface prototype object]
178: expected: FAIL
Added regular file testing/web-platform/meta/WebIDL/idlharness.any.js.ini:
1: [idlharness.any.worker.html]
2: expected:
3: if (os == "android") and fission: [OK, TIMEOUT]
4:
5: [idlharness.any.html]
6: expected:
7: if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/current-realm.html.ini:
1 : [current-realm.html]
2 : [getImageData]
3 : expected: FAIL
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/allow-resizable.html.ini:
1 : [allow-resizable.html]
2 : [APIs without [AllowResizable\] throw when passed resizable ArrayBuffers]
3 : expected: FAIL
4 :
5 : [APIs with [AllowShared\] but without [AllowResizable\] throw when passed growable SharedArrayBuffers]
6 : expected: FAIL
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/builtin-function-properties.any.js.ini:
1 : [builtin-function-properties.any.html]
2 :
3 : [builtin-function-properties.any.worker.html]
4 : expected:
5 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/class-string-interface.any.js.ini:
1 : [class-string-interface.any.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [class-string-interface.any.worker.html]
6 : expected:
7 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/class-string-iterator-prototype-object.any.js.ini:
1 : [class-string-iterator-prototype-object.any.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [class-string-iterator-prototype-object.any.worker.html]
6 : expected:
7 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/class-string-named-properties-object.window.js.ini:
1 : [class-string-named-properties-object.window.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/constructors.html.ini:
1 : [constructors.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/default-iterator-object.html.ini:
1 : [default-iterator-object.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/default-toJSON-cross-realm.html.ini:
1 : [default-toJSON-cross-realm.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/es-exceptions/DOMException-constructor-and-prototype.any.js.ini:
1 : [DOMException-constructor-and-prototype.any.worker.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [DOMException-constructor-and-prototype.any.html]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/es-exceptions/DOMException-constructor-behavior.any.js.ini:
1 : [DOMException-constructor-behavior.any.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [DOMException-constructor-behavior.any.worker.html]
6 : expected:
7 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/es-exceptions/DOMException-custom-bindings.any.js.ini:
1 : [DOMException-custom-bindings.any.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [DOMException-custom-bindings.any.worker.html]
6 : expected:
7 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/es-exceptions/exceptions.html.ini:
1 : [exceptions.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/global-immutable-prototype.any.js.ini:
1 : [global-immutable-prototype.any.worker.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [global-immutable-prototype.any.serviceworker.html]
6 : expected:
7 : if (os == "android") and fission: [OK, TIMEOUT]
8 :
9 : [global-immutable-prototype.any.html]
10 : expected:
11 : if (os == "android") and fission: [OK, TIMEOUT]
12 :
13 : [global-immutable-prototype.any.sharedworker.html]
14 : expected:
15 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/global-object-implicit-this-value-cross-realm.html.ini:
1 : [global-object-implicit-this-value-cross-realm.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/global-object-implicit-this-value.any.js.ini:
1 : [global-object-implicit-this-value.any.worker.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [global-object-implicit-this-value.any.serviceworker.html]
6 : expected:
7 : if (os == "android") and fission: [OK, TIMEOUT]
8 :
9 : [global-object-implicit-this-value.any.html]
10 : expected:
11 : if (os == "android") and fission: [OK, TIMEOUT]
12 :
13 : [global-object-implicit-this-value.any.sharedworker.html]
14 : expected:
15 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/has-instance.html.ini:
1 : [has-instance.html]
2 : prefs: [dom.webidl.crosscontext_hasinstance.enabled:false]
3 : expected:
4 : if (os == "android") and fission: [TIMEOUT, OK]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/interface-object-set-receiver.html.ini:
1 : [interface-object-set-receiver.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/interface-object.html.ini:
1 : [interface-object.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/interface-prototype-constructor-set-receiver.html.ini:
1 : [interface-prototype-constructor-set-receiver.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/interface-prototype-object.html.ini:
1 : [interface-prototype-object.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/invalid-this-value-cross-realm.html.ini:
1 : [invalid-this-value-cross-realm.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/iterator-invalidation-foreach.html.ini:
1 : [iterator-invalidation-foreach.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/iterator-prototype-object.html.ini:
1 : [iterator-prototype-object.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/legacy-callback-interface-object.html.ini:
1 : [legacy-callback-interface-object.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/legacy-factor-function-subclass.window.js.ini:
1 : [legacy-factor-function-subclass.window.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/no-regexp-special-casing.any.js.ini:
1 : [no-regexp-special-casing.any.worker.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [no-regexp-special-casing.any.html]
6 : expected:
7 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/observable-array-no-leak-of-internals.window.js.ini:
1 : [observable-array-no-leak-of-internals.window.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/observable-array-ownkeys.window.js.ini:
1 : [observable-array-ownkeys.window.html]
2 : prefs: [layout.css.constructable-stylesheets.enabled:true]
3 : expected:
4 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/sequence-conversion.html.ini:
1 : [sequence-conversion.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
Removed regular file testing/web-platform/meta/webidl/ecmascript-binding/window-named-properties-object.html.ini:
1 : [window-named-properties-object.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 : [[[OwnPropertyKeys\]\]]
5 : expected: FAIL
Removed regular file testing/web-platform/meta/webidl/idlharness-shadowrealm.window.js.ini:
1 : [idlharness-shadowrealm.window.html]
2 : prefs: [javascript.options.experimental.shadow_realms:true]
3 : [DOMException interface: existence and properties of interface object]
4 : expected: FAIL
5 :
6 : [DOMException interface object length]
7 : expected: FAIL
8 :
9 : [DOMException interface object name]
10 : expected: FAIL
11 :
12 : [DOMException interface: existence and properties of interface prototype object]
13 : expected: FAIL
14 :
15 : [DOMException interface: existence and properties of interface prototype object's "constructor" property]
16 : expected: FAIL
17 :
18 : [DOMException interface: existence and properties of interface prototype object's @@unscopables property]
19 : expected: FAIL
20 :
21 : [DOMException interface: attribute name]
22 : expected: FAIL
23 :
24 : [DOMException interface: attribute message]
25 : expected: FAIL
26 :
27 : [DOMException interface: attribute code]
28 : expected: FAIL
29 :
30 : [DOMException interface: constant INDEX_SIZE_ERR on interface object]
31 : expected: FAIL
32 :
33 : [DOMException interface: constant INDEX_SIZE_ERR on interface prototype object]
34 : expected: FAIL
35 :
36 : [DOMException interface: constant DOMSTRING_SIZE_ERR on interface object]
37 : expected: FAIL
38 :
39 : [DOMException interface: constant DOMSTRING_SIZE_ERR on interface prototype object]
40 : expected: FAIL
41 :
42 : [DOMException interface: constant HIERARCHY_REQUEST_ERR on interface object]
43 : expected: FAIL
44 :
45 : [DOMException interface: constant HIERARCHY_REQUEST_ERR on interface prototype object]
46 : expected: FAIL
47 :
48 : [DOMException interface: constant WRONG_DOCUMENT_ERR on interface object]
49 : expected: FAIL
50 :
51 : [DOMException interface: constant WRONG_DOCUMENT_ERR on interface prototype object]
52 : expected: FAIL
53 :
54 : [DOMException interface: constant INVALID_CHARACTER_ERR on interface object]
55 : expected: FAIL
56 :
57 : [DOMException interface: constant INVALID_CHARACTER_ERR on interface prototype object]
58 : expected: FAIL
59 :
60 : [DOMException interface: constant NO_DATA_ALLOWED_ERR on interface object]
61 : expected: FAIL
62 :
63 : [DOMException interface: constant NO_DATA_ALLOWED_ERR on interface prototype object]
64 : expected: FAIL
65 :
66 : [DOMException interface: constant NO_MODIFICATION_ALLOWED_ERR on interface object]
67 : expected: FAIL
68 :
69 : [DOMException interface: constant NO_MODIFICATION_ALLOWED_ERR on interface prototype object]
70 : expected: FAIL
71 :
72 : [DOMException interface: constant NOT_FOUND_ERR on interface object]
73 : expected: FAIL
74 :
75 : [DOMException interface: constant NOT_FOUND_ERR on interface prototype object]
76 : expected: FAIL
77 :
78 : [DOMException interface: constant NOT_SUPPORTED_ERR on interface object]
79 : expected: FAIL
80 :
81 : [DOMException interface: constant NOT_SUPPORTED_ERR on interface prototype object]
82 : expected: FAIL
83 :
84 : [DOMException interface: constant INUSE_ATTRIBUTE_ERR on interface object]
85 : expected: FAIL
86 :
87 : [DOMException interface: constant INUSE_ATTRIBUTE_ERR on interface prototype object]
88 : expected: FAIL
89 :
90 : [DOMException interface: constant INVALID_STATE_ERR on interface object]
91 : expected: FAIL
92 :
93 : [DOMException interface: constant INVALID_STATE_ERR on interface prototype object]
94 : expected: FAIL
95 :
96 : [DOMException interface: constant SYNTAX_ERR on interface object]
97 : expected: FAIL
98 :
99 : [DOMException interface: constant SYNTAX_ERR on interface prototype object]
100 : expected: FAIL
101 :
102 : [DOMException interface: constant INVALID_MODIFICATION_ERR on interface object]
103 : expected: FAIL
104 :
105 : [DOMException interface: constant INVALID_MODIFICATION_ERR on interface prototype object]
106 : expected: FAIL
107 :
108 : [DOMException interface: constant NAMESPACE_ERR on interface object]
109 : expected: FAIL
110 :
111 : [DOMException interface: constant NAMESPACE_ERR on interface prototype object]
112 : expected: FAIL
113 :
114 : [DOMException interface: constant INVALID_ACCESS_ERR on interface object]
115 : expected: FAIL
116 :
117 : [DOMException interface: constant INVALID_ACCESS_ERR on interface prototype object]
118 : expected: FAIL
119 :
120 : [DOMException interface: constant VALIDATION_ERR on interface object]
121 : expected: FAIL
122 :
123 : [DOMException interface: constant VALIDATION_ERR on interface prototype object]
124 : expected: FAIL
125 :
126 : [DOMException interface: constant TYPE_MISMATCH_ERR on interface object]
127 : expected: FAIL
128 :
129 : [DOMException interface: constant TYPE_MISMATCH_ERR on interface prototype object]
130 : expected: FAIL
131 :
132 : [DOMException interface: constant SECURITY_ERR on interface object]
133 : expected: FAIL
134 :
135 : [DOMException interface: constant SECURITY_ERR on interface prototype object]
136 : expected: FAIL
137 :
138 : [DOMException interface: constant NETWORK_ERR on interface object]
139 : expected: FAIL
140 :
141 : [DOMException interface: constant NETWORK_ERR on interface prototype object]
142 : expected: FAIL
143 :
144 : [DOMException interface: constant ABORT_ERR on interface object]
145 : expected: FAIL
146 :
147 : [DOMException interface: constant ABORT_ERR on interface prototype object]
148 : expected: FAIL
149 :
150 : [DOMException interface: constant URL_MISMATCH_ERR on interface object]
151 : expected: FAIL
152 :
153 : [DOMException interface: constant URL_MISMATCH_ERR on interface prototype object]
154 : expected: FAIL
155 :
156 : [DOMException interface: constant QUOTA_EXCEEDED_ERR on interface object]
157 : expected: FAIL
158 :
159 : [DOMException interface: constant QUOTA_EXCEEDED_ERR on interface prototype object]
160 : expected: FAIL
161 :
162 : [DOMException interface: constant TIMEOUT_ERR on interface object]
163 : expected: FAIL
164 :
165 : [DOMException interface: constant TIMEOUT_ERR on interface prototype object]
166 : expected: FAIL
167 :
168 : [DOMException interface: constant INVALID_NODE_TYPE_ERR on interface object]
169 : expected: FAIL
170 :
171 : [DOMException interface: constant INVALID_NODE_TYPE_ERR on interface prototype object]
172 : expected: FAIL
173 :
174 : [DOMException interface: constant DATA_CLONE_ERR on interface object]
175 : expected: FAIL
176 :
177 : [DOMException interface: constant DATA_CLONE_ERR on interface prototype object]
178 : expected: FAIL
Removed regular file testing/web-platform/meta/webidl/idlharness.any.js.ini:
1 : [idlharness.any.worker.html]
2 : expected:
3 : if (os == "android") and fission: [OK, TIMEOUT]
4 :
5 : [idlharness.any.html]
6 : expected:
7 : if (os == "android") and fission: [OK, TIMEOUT]
It looks like the difference is due to casing, i.e. testing/web-platform/meta/WebIDL/current-realm.html.ini vs testing/web-platform/meta/webidl/current-realm.html.ini. Interestingly, all of the problematic files are .ini.
Specifications
- Platform: macOS 13.4 (22F66)
- Version: jj built at d01ecc5c46043351ea0a3ab982bb2fd606d81ea3
We have this TODO about that: https://github.com/martinvonz/jj/blob/d01ecc5c46043351ea0a3ab982bb2fd606d81ea3/lib/src/working_copy.rs#L60-L61
I'm not sure if there's a better solution. Since we have recently talked about making TreeState more of a pure cache than it currently is, maybe we should not store the information there (as that TODO suggests).
This is not a fix, but a comment to https://www.mgaudet.ca/technical/2023/11/23/exploring-jujitsu-jj suggested a workaround of creating a case-sensitive APFS volume, which is cheap on a Mac. Unfortunately, it seems like many Mac-native apps are untested and buggy on a case-sensitive volume, so you can't live your life on such a volume.
I'm unsure about Windows, the analogue might require repartitioning there.
I've added a simple test case for this. I also added the Option<PathBuf> to FileState, but I'm not sure when it would get set other than None :) I'll have more of a poke, but thought I might as well share this now for comment.
https://github.com/martinvonz/jj/commit/ecfbb2531df39d369903eb4d6922c2141400b3dc
I've added a simple test case for this. I also added the
Option<PathBuf>toFileState, but I'm not sure when it would get set other thanNone:) I'll have more of a poke, but thought I might as well share this now for comment.
I think I was thinking that we would set it to the actual path on disk if it was different from the RepoPath but I don't think I spent much time thinking about how it would work in detail.
Also note this comment from above:
Since we have recently talked about making TreeState more of a pure cache than it currently is, maybe we should not store the information there (as that TODO suggests).
Maybe it's better to walk the directory tree on disk and then check what's already in the base tree to decide which path to write to the new tree. For example, if we see a file called "foo" on disk and there's an existing file called "Foo" in the tree, we update that path instead (but only if we know we're on a case-insensitive file system, of course).
We will also have to decide what to do if there are multiple paths with different case in the tree. We will probably want to make snapshotting consistent with what we do on checkout. For example, if there's both "Foo" and "foo" in the tree (commit), we might decide that we put the "Foo" path in the working copy. Then, when we later snapshot the working copy, we should make sure we write a new tree where we update the "Foo" path with any changes to the "foo"/"Foo" file on disk.
Thanks for the comment. I suspect that "full" support for case-insensitive file-systems will be a multi-pronged effort. In this first instance I'd be very happy with just allowing a tree to be checked out and considered clean, allowing the rest of the repo to be used. That fact things might fall apart if you try and change such files could be a followup.
I was wondering if this use-case might be able to avoid a new configuration option - after writing the file "FOO/bar" we could ask the file-system what the actual name is, and when told it's actually "foo/bar" (due to that directory already existing) we could adjust things accordingly.
I'm very new to jj, and even newer to the code and my grasp of the terminology is going to be tenuous, so apologies for any confusion, but:
Maybe it's better to walk the directory tree on disk and then check what's already in the base tree to decide which path to write to the new tree. For example, if we see a file called "foo" on disk and there's an existing file called "Foo" in the tree, we update that path instead
In an initial checkout we obviously can't look at the file-system before any operations. In my example, I suspect that whichever file is processed first out of "FOO" and "foo", the file-system will match that case. Then when the second file attempts to create "foo/bar" (or "FOO/bar") it will succeed, but the "foo" part will have an unexpected case.
Does that make sense? Is that just re-phrasing what you were saying, or am I missing the bigger picture?
Maybe it's better to walk the directory tree on disk and then check what's already in the base tree to decide which path to write to the new tree.
I've looked at this some more and dug into how the tree-walking works for a snaphot. Like ecfbb25, consider snapshotting the pathological:
- repo paths: A/b/C, a/B/d
- which was written into a wc as: A/b/C, A/b/d
IIUC, as TreeState::visit_directory() is visiting A/b/d, it sets up the state for that file in the tree here - but all we have there is the path to the file (ie, the A/b/d) and not the path we need, a/B/d. Because this path is a flat key from the root of the repo, I can't see an efficient way of discovering a/B/d given just A/b/d (ie, an efficient way to "walk up" the tree to resolve each of the path components individually)
And without that we have the current situation: jj believes a/B/d is missing while A/b/d is a new untracked file.
Is my above understanding correct? Any suggestions for how I might proceed with this experiment?
- repo paths: A/b/C, a/B/d
- which was written into a wc as: A/b/C, A/b/d
Good example!
I actually wonder if it's even correct to write the tree repo paths to the working copy in that way. Maybe we should skip the entire a/ directory since we had already written the A/ directory (assuming we let the first entry in lexicographical order win). It seems odd to combine them into a single directory. That would likely break builds. Of course, completely skipping a/ would also break the build, but at least the build targets inside A/ would be more likely to work.
What do you think about skipping entries in a Tree object if we've already visited an entry that maps to the same path? But how do we even know if they map to the same path? Does that depend on the file system? I know the unicode rules for lowercasing are complicated. Hmm, a quick web search led to this FAQ, which seems to say that HFS+ and APFS do it differently. So I suppose we have to open the file for writes and then ask the file system for the name to see if it matches another path in the tree?
Btw, do we want to suppose only case-insensitive file systems that are at least case-preserving (I think some filesystems will give you a file called "A" if you write a file called "a")? Maybe it's not much work to support both.
Do you have thoughts on when we would detect a case insensitive file system? I'm thinking we would do that when we initialize the TreeState and then store a flag in it.
Might be worth looking into how other tools (e.g. Git and Mercurial) handle these things. It seems pretty thorny.
What do you think about skipping entries in a
Treeobject if we've already visited an entry that maps to the same path?
I think there are 2 questions here - how should we write a working-tree vs how do we snapshot an existing working tree. In the original example we are snapshotting a working-tree created by git. ie, we are walking the file-system and discover the directory testing/web-platform/meta/WebIDL without knowing the repo has testing/web-platform/meta/WebIDL and testing/web-platform/meta/webidl
Might be worth looking into how other tools (e.g. Git and Mercurial) handle these things.
When it comes to writing the working tree, git does what my example above will do - ie, create all the files in a working-tree that doesn't exactly match the repo. Things will quickly fall apart if there's A/b and a/b, but I think that's OK (ie, such a tree would simply not be supported on a case-insenstive fs)
In my experiments, that's also what jj does - IOW, jj and git seem to agree on what a working-tree looks like when creating it. It's been years since I used hg, but I'd be super surprised if it doesn't do the same thing when writing a tree, simply because that's what a naive implementation does automatically, and also because gecko-dev works with hg on mac/windows and any other behaviour (ie, not checking out the mis-matched files etc) would not.
Btw, do we want to suppose only case-insensitive file systems that are at least case-preserving
I think it's probably fine to assume case-preserving these days (even fat32 is iiuc), but I don't see how that changes things though - case-preserving means the first-writer wins but the problem is the "losing" writers. Non case-preserving would just mean that first-writer also becomes a "loser" in this regard, but we need to deal with losers anyway.
Do you have thoughts on when we would detect a case insensitive file system?
auto-detecting would be fantastic. I'd be happy to lean into this once things are in-place to work with it.
auto-detecting would be fantastic. I'd be happy to lean into this once things are in-place to work with it.
The most reliable and FS-independant way I know of to detect such systems (and all future case insensitive FS) is to create a file/or dir and check it's existence out in another casing:
std::fs::create_dir("<path to .jj>/CaSe-deTECtion/")?;
assert_eq!(Path::new("<path to .jj>/cAsE-DEtectTION").try_exists(), Ok(true))
If we store this as metadata we'll need to recheck it periodically, since a jj repo could be moved to another FS. The simplest way for that is probably to save our current absolute path and if it changes, recompute the case sensitivity ?
Auto-detection is definitely part of the story, but each filesystem has its own case-folding and normalization rules on top (and the normalization rules apply even on some case-sensitive filesystems).
For example, as documented here, APFS uses Unicode NFD for the most part, but there are exceptions. There's also some weirdness around zero-width characters that I don't quite recall at the moment.
And this applies to case-sensitive APFS as well.
We could probably handle simple cases with a flag, but in general this is a hugely difficult problem.
(and the normalization rules apply even on some case-sensitive filesystems).
I guess those case-sensitive file-systems will have the same problem - if walking the fs provides paths which aren't identical to the original tree, jj will get confused about what files are missing/new etc?
but in general this is a hugely difficult problem.
Indeed, but is there a way to make small steps here? If we scale the problem down to, say "just ascii case normalization" (or more abstractly "support reading a working tree created by git") we'd probably solve the vast majority of these cases in the wild without compromising the ability to go further in the future?
Indeed, but is there a way to make small steps here? If we scale the problem down to, say "just ascii case normalization"
That's fine with me.
(or more abstractly "support reading a working tree created by git") we'd probably solve the vast majority of these cases in the wild without compromising the ability to go further in the future?
Matching Git is a bit more limiting but probably fine. Do you want to investigate what Git does? We would also need to be able to write a working copy ourselves, and I suppose we would want Git to be able to read the working copy after we write it then?
Matching Git is a bit more limiting but probably fine.
When I said "matching git" I meant in a very loose way. git doesn't necessarily do the best thing, just a minimal "good enough" thing. For example, doing git log A/b (matching the file-system but not the repo) gives zero log entries, whereas git log a/B (ie, matching the repo but not the file-system) shows the actual log. I don't think we should necessarily aspire to that level of "matching git"
Also, to be very explicit, I really don't think we should see this capability as a "feature" but instead as a work-around. I've never seen a case-mismatch like this which was intentional, but instead it was accidentally introduced and likely to be "fixed" at some point for (some revisions of) that tree anyway. So IMO it's OK for jj to not aim for the best thing to do in some of these cases, but instead just avoid an obviously bad thing.
Do you want to investigate what Git does?
git has some interesting behaviour when adding files. eg:
let's add A/b/C to a repo
$ git init foo && cd foo
$ mkdir -p A/b & touch A/b/C
$ git add A
$ git ls-files
A/b/C
$ git ci -am 1
all good. Now let's nuke the files and try to add a/B/c
$ rm -rf A
$ mkdir -p a/B & touch a/B/d
$ # verify the file-system has the case we specified
$ find a
a
a/B
a/B/d
$ # Add the new file
$ git add a/B/d
$ # restore the file we removed.
$ git restore A
$ # The repo
$ git ls-files
A/b/C
A/b/d
Note that git created the new file with a path that matches the repo and not the file-system.
$ # The file-system
$ find a
a
a/B
a/B/C
a/B/d
git status is clean. Further experiments show that I can manually make a working tree with all case combinations and git doesn't care - git status is always clean regardless of case-mismatches here.
If I trick git into allowing mixed-case tree entries via, eg, merging a branch, git does get a little confused:
eg, create the main branch:
$ git init foo && cd foo
Initialized empty Git repository in /private/tmp/foo/foo/.git/
$ touch README && git add README && git ci -am README
create a branch with one variation of case:
$ git co -b b1
Switched to a new branch 'b1'
$ mkdir -p A/b && touch A/b/C && git add A/b/C && git ci -am "first c"
another branch with another variation, then merge the branches:
$ git co -b b2 main
Switched to a new branch 'b2'
$ mkdir -p a/B && touch a/B/c && git add a/B/c && git ci -am "second c"
$ git merge b1
Merge made by the 'ort' strategy.
A/b/C | 0
1 file changed, 0 insertions(+), 0 deletions(-)
create mode 100644 A/b/C
git now has two variations of the same file:
$ git ls-files
A/b/C
README
a/B/c
$ git st
On branch b2
nothing to commit, working tree clean
However, touching the file-system doesn't impact git:
$ echo new > a/B/C
$ git st
On branch b2
nothing to commit, working tree clean
(ie, presumably you'd expect one of the variations of C to be dirty, but the tree is clean. It's not clear if this is intentional or an implementation accident, but regardless, there's no objectively "correct" thing to do.)
eg, if I check-out that tree:
$ rm -rf A
$ ls
README
$ git co .
Updated 2 paths from the index
git says it updated 2 paths, but it didn't - obviously there's only 1 c file on disk.
We would also need to be able to write a working copy ourselves, and I suppose we would want Git to be able to read the working copy after we write it then?
My experiments show that git really doesn't care what the case of the file-system is vs the tree. Even if git and jj would end up creating different file-systems but where the only difference is the case, then git will be happy. Or if not happy, at least no more confused than it can make itself.
I'll try and get some insight into what git actually does with the core.ignorecase setting too, but I suspect I'll find that code-base fairly intractable.
I'll try and get some insight into what git actually does with the
core.ignorecasesetting too, but I suspect I'll find that code-base fairly intractable.
heh, I was wrong. This comment describes what I saw, and the implementation is here
They also probe the file-system to automatically set core.ignorecase
FWIW, https://hg.mozilla.org/mozilla-central/rev/3b6c10d28209 just landed, which means the current tip of gecko no longer exhibits this problem (although obviously it may come back at any time and still exists for older revisions). It might be worth updating the title of this issue to be more generic?
I guess those case-sensitive file-systems will have the same problem - if walking the fs provides paths which aren't identical to the original tree, jj will get confused about what files are missing/new etc?
Yeah, any kind of normalization needs to be accounted for. (From what I remember, Git wasn't quite perfect either.)
Indeed, but is there a way to make small steps here? If we scale the problem down to, say "just ascii case normalization" (or more abstractly "support reading a working tree created by git") we'd probably solve the vast majority of these cases in the wild without compromising the ability to go further in the future?
Yeah, though even, say codebases with accents, which I'm going to guess are pretty common, are going to run into Unicode normalization issues. For example, if a filename is stored in Unicode NFC, on macOS it'll get written out as NFD -- and we'd need a way to match up the names internal to jj with the ones on disk. So it's worth at least thinking more holistically about the problem, even if the initial solution chosen isn't complete.