edfcore 0.5.12 → 0.5.14
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 +42 -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 = "0.5.
|
|
112
|
+
export declare const VERSION = "0.5.14";
|
|
113
113
|
//# sourceMappingURL=constants.d.ts.map
|
package/dist/constants.js
CHANGED
package/docs/CHANGELOG.md
CHANGED
|
@@ -6,6 +6,48 @@ 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
|
+
## 0.5.14
|
|
10
|
+
|
|
11
|
+
- **Added** the CLI's default print cap driven at all four places it is applied.
|
|
12
|
+
`cli-limit-default.test.ts` covers the NUMBER — twenty, and the four documents that promise it —
|
|
13
|
+
and its own docblock names the failure that made `DEFAULT_ITEM_LIMIT` a constant in 0.4.390: the
|
|
14
|
+
literal `20` at four separate call sites, "`header` printing twenty diagnostics while
|
|
15
|
+
`events --list` prints fifty, with both pages still saying twenty". It then exercised one of the
|
|
16
|
+
four. The three the sentence is about had never been run with the flag left off.
|
|
17
|
+
- They are not reachable by accident. Each needs more than twenty of one kind of diagnostic in one
|
|
18
|
+
particular array: `header.diagnostics`, then `timeline.diagnostics` under "From the record
|
|
19
|
+
probes:" — a second application of the same local, added in 0.3.94 — and `report.diagnostics`
|
|
20
|
+
through `formatValidationReport`, which is a different formatter with its own `maxItems`. The
|
|
21
|
+
fixtures are built for it: thirty signals declaring `physicalMinimum == physicalMaximum`, thirty
|
|
22
|
+
annotation signals each carrying a malformed slot-0 TAL that both record probes see, and a
|
|
23
|
+
malformed TAL in every one of forty records.
|
|
24
|
+
- Every count is compared against the cap the CLI was observed to apply rather than against a
|
|
25
|
+
literal, so this file names no number of its own and the two files move together. An explicit
|
|
26
|
+
`--limit 3` is then driven through all four as well, which is what separates a site that reads
|
|
27
|
+
`args.limit` from one that fell through to the constant.
|
|
28
|
+
|
|
29
|
+
## 0.5.13
|
|
30
|
+
|
|
31
|
+
- **Added** a `NaN` budget driven through every entry point that reads one. `options.ts` exists
|
|
32
|
+
because `maxMaterializeBytes` is typed `number`, which admits `NaN`, and `NaN` arrives without
|
|
33
|
+
anyone writing it — `Number(process.env.EDF_BUDGET)`, an absent query parameter, a missing JSON
|
|
34
|
+
key. Its docblock states the reach the guard needs: the option is resolved in six modules and read
|
|
35
|
+
raw and handed on in two more, and "a guard that only one of the eight applies is not a guard".
|
|
36
|
+
- Nothing checked the eight. `options.test.ts` checks the resolver in isolation, and
|
|
37
|
+
`budget-boundary.test.ts` enumerates the five sites that compare a REQUIREMENT against the budget
|
|
38
|
+
— a different set, reached only by a request large enough to refuse. A `NaN` is refused earlier
|
|
39
|
+
than that, by the resolver, in modules that list names: `record-index.ts` compares nothing, and
|
|
40
|
+
`biosemi.ts` resolves nothing.
|
|
41
|
+
- The eight are now read out of `src/` and split by what they do with the option, with the counts
|
|
42
|
+
checked against the sentence in the docblock that states them. Each is then driven through its own
|
|
43
|
+
public entry point — `readRecordBytes`, `readRecords`, `decodeDigital`, `toPhysical`,
|
|
44
|
+
`readEnvelope`, `buildRecordIndex`, `validateRecording` on both its branches, and `readTriggers`,
|
|
45
|
+
which owns no resolver and borrows the scan chunker's.
|
|
46
|
+
- The eighth is stated rather than asserted away: `cachedSource` hands the option to the source it
|
|
47
|
+
wraps, so a `NaN` survives it and is refused by the first resolving module downstream. That is the
|
|
48
|
+
design — a cache is not the layer that owns a materialisation budget — and it is now a fact the
|
|
49
|
+
suite records instead of a gap in it.
|
|
50
|
+
|
|
9
51
|
## 0.5.12
|
|
10
52
|
|
|
11
53
|
- **Added** the six mistakes `AGENTS.md` tells a code generator to avoid, run. That file carries two
|
package/package.json
CHANGED
package/src/constants.ts
CHANGED