edfcore 0.4.353 → 0.4.355
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/dist/constants.d.ts +1 -1
- package/dist/constants.js +1 -1
- package/docs/CHANGELOG.md +29 -0
- package/package.json +1 -1
- package/src/constants.ts +1 -1
package/dist/constants.d.ts
CHANGED
|
@@ -109,5 +109,5 @@ export declare const SIGNAL_FIELD_BLOCK_OFFSETS: {
|
|
|
109
109
|
readonly reserved: 224;
|
|
110
110
|
};
|
|
111
111
|
/** Published package version. Kept in sync with package.json by a test. */
|
|
112
|
-
export declare const VERSION = "0.4.
|
|
112
|
+
export declare const VERSION = "0.4.355";
|
|
113
113
|
//# sourceMappingURL=constants.d.ts.map
|
package/dist/constants.js
CHANGED
package/docs/CHANGELOG.md
CHANGED
|
@@ -6,6 +6,35 @@ alone does not tell you whether you were affected.
|
|
|
6
6
|
edfcore is pre-1.0. Patch releases have carried behaviour changes where the old behaviour was a
|
|
7
7
|
defect; those are called out below.
|
|
8
8
|
|
|
9
|
+
## 0.4.355
|
|
10
|
+
|
|
11
|
+
- **Fixed** a test of mine that could time out. The signature check added in 0.4.354 spawned `tsc`
|
|
12
|
+
inside an `it`, which takes a couple of seconds alone and rather more under a loaded suite —
|
|
13
|
+
past the default per-test timeout. It passed every run alone and failed the first time the whole
|
|
14
|
+
suite ran it.
|
|
15
|
+
- Now compiled once at module scope, which is what `doc-snippets-compile.test.ts` next door
|
|
16
|
+
already does with its own compiler run. Collection time is not on a per-test clock, so the work
|
|
17
|
+
happens where it is not being timed, and nothing about the check weakens.
|
|
18
|
+
|
|
19
|
+
## 0.4.354
|
|
20
|
+
|
|
21
|
+
- **Added** a compile check over every signature the reference pages print. Four pages open each
|
|
22
|
+
function with a `ts` fence holding its declaration and nothing else — thirty-odd across
|
|
23
|
+
`api-primitives`, `api-reading`, `api-sources` and `api-validate` — and every one is hand-typed.
|
|
24
|
+
They are the first thing a reader sees for a function and what they write their call against.
|
|
25
|
+
- `doc-snippets-compile.test.ts` compiles the fences that are complete programs and skips these: a
|
|
26
|
+
bare declaration imports nothing, so it never matched the filter that finds runnable examples.
|
|
27
|
+
The most load-bearing line on each reference page was the one nothing compiled.
|
|
28
|
+
- Assignability is checked in BOTH directions, which is the difference between "the documented
|
|
29
|
+
shape is usable" and "the documented shape is the real one" — one direction alone accepts a page
|
|
30
|
+
that widens a parameter or narrows a return, and a widened parameter describes a function that
|
|
31
|
+
does not exist. Signatures are extracted per fence rather than by scanning the page, because a
|
|
32
|
+
lazy match for the parameter list's closing paren runs past it into the prose: `Promise<{ … }>`
|
|
33
|
+
has no `):` to stop at and the next parenthetical sentence does.
|
|
34
|
+
- The types a declaration mentions are resolved through `barrel-types.ts` rather than a hard-coded
|
|
35
|
+
list. The first version used a list and missed four, which is the failure that helper was
|
|
36
|
+
extracted to stop repeating.
|
|
37
|
+
|
|
9
38
|
## 0.4.353
|
|
10
39
|
|
|
11
40
|
- **Added** a cross-check between the two places `openEdf`'s cost is published. `large-files.md`
|
package/package.json
CHANGED
package/src/constants.ts
CHANGED