smodusermc's picture
|
download
raw
4.02 kB
# Offline games archive (original files)
These are the original files of the locally hosted games and web ports listed at https://garbsoftball.com/g, stored under the site's own folder layout. The bucket was made public at the owner's request. Rights stay with the games' owners.
## How to play (read this first)
**Do not open a game's `index.html` straight from disk.** Browsers block a page opened as `file://` from loading its game data, so the game stops at its loading screen (Subway Surfers, for example, stays at "0% of 35MB Loading..."). The games must be served over `http://`, as the original site did. `serve.py` does this and needs only Python 3.8 or newer:
1. Download `serve.py` from the top of this bucket (Windows: also `serve.bat`; macOS: `serve.command`).
2. In the folder where you saved it, run one of:
- `python3 serve.py --get "subway surfers"` (Windows: `py serve.py --get "subway surfers"`). This downloads that game's files from this bucket, checks each by SHA-256, starts a local server and opens the game. Use a name, part of a name, or the id from `_reports/inventory.json`.
- `python3 serve.py`: serves everything already downloaded and opens a searchable game list.
3. To download many games at once, use the `hf` CLI, then run `python3 serve.py`:
- everything (about 79 GB): `hf buckets sync hf://buckets/smodusermc/offline-games/site ./site`;
- one folder: `hf buckets sync hf://buckets/smodusermc/offline-games/site/games/subway-surfers ./site/games/subway-surfers`.
Some games also use shared folders, for example `site/extra/` for emulator cores; `--get` fetches everything a game lists.
Some games load a public library from a CDN (for example the Ruffle Flash player from unpkg.com). `--get` also downloads the stored copies of those libraries (`_vendor/`), and `serve.py` points the game page at them, so those games work offline too.
## Status
- **All 604 catalog entries have been processed:**
- 297 captured (32,296 unique files, 79.28 GB of game files);
- 300 skipped, because the game itself is hosted on another site or embedded from one;
- Geometry Dash left out, because it was delivered separately;
- 6 not capturable or deferred (the reasons are in the review).
- The bucket holds about **79.39 GB**, including reports and evidence.
- **Archive-wide review:** `_verification/FINAL-REVIEW.md`. Per-entry table: `_verification/final-review-entries.csv`.
- **Latest batch report:** `_verification/LATEST.md` (batch 6). Every batch report is at `_verification/batches/<batch>/README.md`.
## Layout
- `site/`: the site's files at their original paths (`site/games/<game>/…`, `site/gamefile/…`, `site/extra/…`).
- `_vendor/`: public runtime libraries that some games load from CDNs, stored so the builds could be tested with outside network blocked.
- `_reports/`: `inventory.json` (every catalog entry and its status) and `games/<id>.json` (every stored file, with size and SHA-256).
- `_verification/`: batch reports, browser-run results, screenshots and evidence.
## Using the files
- These are web builds that expect to be served over HTTP from the site root (paths like `/games/…`): use `serve.py` (above) or any static web server that serves `site/` as the root with correct MIME types (for example `application/wasm`). They are not file:// or single-file releases.
- Some builds need cross-origin isolation, as on the live site: `Cross-Origin-Opener-Policy: same-origin` and `Cross-Origin-Embedder-Policy: credentialless`.
- Games that load a library from a CDN need that URL either reachable or routed to the copy in `_vendor/`.
- Every file in every captured entry's manifest was read back from this bucket and matched by SHA-256 and size: most at capture time, and the 74 earliest entries (8,575 files) in a final read-back (`_verification/final-readback.json`).
- Browser checks confirm start-up, menus and some gameplay only; none is a full playthrough. Known source-side gaps and test-machine limits are listed in the review.

Xet Storage Details

Size:
4.02 kB
·
Xet hash:
cdf3a643352b2a74f843b491eea68e53e6cdaa429013f347c79e5ba2449802ac

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.