Roads & highways highways-conus · 2026q2r2
community · highways · in the live catalog · history
Experimental / Amateur-Built use only. Not certified. This data is provided as-is with no warranty and any use of the data is at your own risk.
Checks
pass 8 warn 0 fail 1 not run 6
- fail PACK-002 Attribution and license are recorded in pack_meta: missing: license
- not run PACK-004 Upstream sources are recorded: Retro record: this edition was published before pack-report/1 existed, and the upstream inputs it was built from were not recorded then and cannot be recovered from the pack.
- not run ROW-002 Row counts against the previous edition: Retro record: no earlier edition is listed, so there is nothing to compare against. Editions published before the ones the manifest lists were pruned and are not known here.
- not run ROW-003 No table of the previous edition has disappeared: Retro record: no earlier edition is listed, so there is nothing to compare against. Editions published before the ones the manifest lists were pruned and are not known here.
- not run OSM-001 Feature counts per region against the previous edition: Retro record: no earlier edition is listed, so there is nothing to compare against. Editions published before the ones the manifest lists were pruned and are not known here. This edition's per-region counts are recorded for the next one.
- not run OSM-002 Feature counts per state or province against the previous edition: Needs state and province boundaries; planned for catalogue v2 (TV7). OSM-001 bins by region meanwhile.
- not run OSM-005 Coverage map image: Rendered by the report site (TV5).
- record is retro, not build: only a build record can clear the gate
- PACK-002 failed: missing: license
- ROW-003 is blocking and did not run: Retro record: no earlier edition is listed, so there is nothing to compare against. Editions published before the ones the manifest lists were pruned and are not known here.
- OSM-001 is blocking and did not run: Retro record: no earlier edition is listed, so there is nothing to compare against. Editions published before the ones the manifest lists were pruned and are not known here. This edition's per-region counts are recorded for the next one.
Show:
| Check | What it checks | Measured | Expected | Cannot catch | ||
|---|---|---|---|---|---|---|
| fail | PACK-002 | Attribution and license are recorded in pack_meta | missing: license | both non-empty | A license tag or attribution that is present but wrong for the data. | |
| not run | PACK-004 non-blocking | Upstream sources are recorded | — | at least one source, each with a non-empty edition and a 64-hex sha256 | A recorded source that is not the one the build actually used, and an edition or sha256 that is well formed but wrong. | Note: Retro record: this edition was published before pack-report/1 existed, and the upstream inputs it was built from were not recorded then and cannot be recovered from the pack. |
| not run | ROW-002 non-blocking | Row counts against the previous edition | — | Provisional threshold: no table moves more than 10%; a larger move needs a build note | A change that keeps counts steady (rows replaced by wrong rows), any defect already present in the previous edition, and a loss under the threshold (for obstacles at 10%, about 65,000 rows). | Note: Retro record: no earlier edition is listed, so there is nothing to compare against. Editions published before the ones the manifest lists were pruned and are not known here. |
| not run | ROW-003 | No table of the previous edition has disappeared | — | every table in the previous edition is still present, unless schema_version changed | A table dropped together with a schema_version bump that should not have been; the bump is taken as the declaration that the drop is intended. | Note: Retro record: no earlier edition is listed, so there is nothing to compare against. Editions published before the ones the manifest lists were pruned and are not known here. since AER-4102 |
| not run | OSM-001 | Feature counts per region against the previous edition | — | Provisional threshold: no region with 100+ features goes to zero (fail); no region drops more than 10% (warn) | A loss smaller than a region (one state inside a multi-state region), and features that are present but wrong. | Note: Retro record: no earlier edition is listed, so there is nothing to compare against. Editions published before the ones the manifest lists were pruned and are not known here. This edition's per-region counts are recorded for the next one. since makerplane-data#17 |
| not run | OSM-002 non-blocking | Feature counts per state or province against the previous edition | — | no state goes to zero | Losses inside a state. | Note: Needs state and province boundaries; planned for catalogue v2 (TV7). OSM-001 bins by region meanwhile. |
| not run | OSM-005 non-blocking | Coverage map image | — | an image a person can look at | Anything; it is for a person to read, not a test. | Note: Rendered by the report site (TV5). |
| pass | PACK-001 | pack_meta is readable and names a known kind | id=highways-conus kind=highways cycle=2026q2r2 schema_version=1 | pack_meta present; kind is one the device knows | A pack_meta that is well formed but names the wrong id or cycle for the bytes inside. | |
| pass | PACK-003 | Origin label is federal, community or user-supplied | community (defaulted by kind highways, not read from the pack) | one of federal / community / user-supplied | Whether the label matches where the data really came from. For a kind with a default label (ORIGIN_BY_KIND in packtools/report.py) the label comes from that table, not from the builder or the pack, so the check is tautological and always passes; measured says which. With --origin it shows only that the label given is one of the three. | since AER-699 |
| pass | ACC-001 | Device acceptance: pyEfis's own reader opens the pack and answers a fixed query set | opened; 5/5 queries returned data | the reader opens the pack and every query returns data | Wrong values the reader accepts. It shows the device can open and query these bytes, not that what it returns is right. The queries cover fixed points only: US kinds, 5 CONUS metros (Denver, Seattle, Atlanta, Los Angeles and Chicago); airports (Canada), Toronto and Vancouver; water, Puget Sound, Lake Michigan and Lake Ontario; navaids, idents DEN, SEA, LAX; procedures, airways V23, J80, V1; terrain, one native tile or the coarsest mosaic level. So it cannot catch a missing Alaska, Hawaii or Puerto Rico, or any state or province that holds none of those points. | |
| pass | SQL-001 | SQLite structure is sound (PRAGMA quick_check) | ok | ok | Wrong data in a structurally sound file, and an index whose contents do not match its table: quick_check does not verify index contents, which is SQLite's documented difference from integrity_check. IDX-001 covers the R-trees. | |
| pass | ROW-001 | No data table is empty | 0 | 0 empty tables, apart from those declared empty by design for the kind (water: waterway_lines) | A table that has rows but is missing most of them. | |
| pass | IDX-001 | Every spatial index (R-tree) indexes every row of its table | highway_rtree 909,542 = highway_lines 909,542, rtreecheck ok | each declared R-tree present, its row count equal to its base table's, and rtreecheck() ok; no undeclared R-tree | An index that holds the right number of rows with wrong boxes or ids. rtreecheck tests the tree's internal consistency, not that each box encloses its feature. | since AER-4102 |
| pass | OSM-004 | Every feature's bounding box is well formed | 0 | 0 bad boxes | A well-formed box around wrong geometry. | |
| pass | OSM-003 | highway_lines columns agree with the declared schema_version | declared 1, columns say 1 | declared schema_version equals the one the column set implies | A column whose meaning changed without its name changing. | since AER-1715 |
Black-box check (Auspex)
Not run for this edition.
Row counts
Rows per table, no previous edition was compared.
| Table | Rows | Previous | Change |
|---|---|---|---|
| highway_lines | 909,542 | — | — |
Identity
| Pack sha256 | 484479ecfe344171030cbda7b587e23af2c24e030f6ae1f038f54ba5aeeb0a57 |
|---|---|
| Size | 115,793,920 bytes |
| Content schema | 1 |
| Record | retro · raw JSON sha256 bb7140f4e28f2af1 — matches the report_sha256 in the signed catalog |
| Built | · · no CI run recorded |
| From | makerplane-data — |
| Compared against | none |
Sources
No upstream sources are recorded for this edition (see PACK-004).
Known issues
Issues found after this edition was built are not listed here yet. Until they are, see the makerplane-data issues.
What this report is and is not
These are the checks we ran on this pack, what each one measured, and what each one cannot catch. It is evidence, not a warranty and not a certification. No authority has reviewed or approved this data or these checks. Passing them does not mean the data is correct, complete or current. Results marked as coming from Auspex come from a proprietary AI test engine whose method is not published; each of its findings is stated so you can check it yourself. This data is for Experimental / Amateur-Built use. This data is provided as-is with no warranty and any use of the data is at your own risk. Verify against official publications.
Check it yourself
packtool report --check highways-conus-2026q2r2.pack
Re-runs every open check on the pack you downloaded and compares the result with this record. How to do it.