Skip to content

fix(providers): promote Oracle Cloud and Nextcloud to stable, show Blomp in the production grid - #577

Open
axpnet wants to merge 3 commits into
mainfrom
fix/oracle-cloud-promote-stable
Open

fix(providers): promote Oracle Cloud and Nextcloud to stable, show Blomp in the production grid#577
axpnet wants to merge 3 commits into
mainfrom
fix/oracle-cloud-promote-stable

Conversation

@axpnet

@axpnet axpnet commented Aug 9, 2026

Copy link
Copy Markdown
Member

Why

Provider visibility flags had drifted from reality, and each drift was telling users something untrue.

Oracle Cloud showed the red "Experimental provider" banner, which says the service may be unavailable and that connections can fail through no fault of the user. The flag behind it, ProviderConfig.stable, means "not fully tested". Oracle arrived with the v1.5.0 provider batch and was never tested: when Alibaba OSS and Tencent COS were promoted in 5272f99, Oracle was left behind rather than found broken.

Nextcloud carried stable: false, // Not tested yet since the entry was written, so the GUI warned that Nextcloud "is not stable yet" while ROADMAP.md lists it among the supported presets. Both could not be true.

Blomp was promoted to production on 2026-07-28 in 798811b, with the registry recording stable: true and the live verification. The protocol grid never followed: it still wrapped Blomp in import.meta.env.DEV under a comment claiming it is "hidden in production until storage proxy 403 is resolved (stable: false in registry)", which the registry has contradicted since that release.

Verification

Both promotions were tested against live accounts before the flag was flipped. Same battery, run end to end, with every test artifact removed afterwards.

Operation Oracle Cloud (eu-milan-1, S3 compat) Nextcloud (self-hosted behind nginx)
connect 398 ms 934 ms, DAV endpoint auto-detected from the bare host
list, mkdir, stat pass pass
quota (df) not supported, see below pass, unmetered total via OCS
upload 64 KB remote sha256 matches local remote sha256 matches local
download roundtrip byte identical byte identical
rename, server-side copy pass pass
upload 32 MB pass in 54 s, sha256 matches (multipart) pass in 56 s, sha256 matches
public share link not applicable created, and 404 after the delete
recursive size scan pass, s3-list-recursive not applicable
delete, recursive delete pass pass

Oracle's df returns Operation not supported: storage_info, which is normal for S3 backends with no quota API and is exactly what the manual "Total storage" field exists for. It is not a reason to warn a user about the service.

Nextcloud listing was cross-checked against a second, independent Nextcloud (Tab.digital) to separate server quirks from client behaviour. The self-hosted box exposes a phantom remote.php collection at every path depth, which is an nginx rewrite on that server: the second Nextcloud lists clean.

Local gate on the branch: tsc --noEmit clean, vitest run 80 files and 718 tests passing, re-run after each commit.

Still flagged

DigitalOcean Spaces and S3Drive keep stable: false. They are untested, not known broken, and promoting them without a live run would repeat the mistake this series fixes.

Not in this PR

The banner copy itself deserves a rewrite. It reads "{provider} is not stable yet and the service may be unavailable", which blames the third party, while the flag only records that we have not verified it. That is a copy change across every locale and belongs in its own PR.

axpnet and others added 2 commits August 9, 2026 16:32
Oracle Cloud shipped with the v1.5.0 provider batch and kept `stable: false`
ever since, which renders the red "Experimental provider" banner in Quick
Connect and tells the user the service may be unavailable. The flag means
"not fully tested" (`ProviderConfig.stable`), and nobody had tested it: when
Alibaba OSS and Tencent COS were promoted in 5272f99, Oracle was left
behind rather than found broken.

Verified against a live OCI account (eu-milan-1, S3 compatibility endpoint):
connect 398ms, list, mkdir, stat, 64 KB upload with a matching remote
sha256, byte-identical download roundtrip, rename, server-side copy, 32 MB
multipart upload with a matching sha256, recursive size scan, delete and
recursive delete. The bucket was left exactly as found.

The one unsupported call is `df` (`storage_info`), which is normal for S3
backends with no quota API and is what the manual "Total storage" field
exists for. It is not a reason to warn the user about the service.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Blomp was promoted to production on 2026-07-28 (798811b): the registry
carries `stable: true` and a comment recording the live verification, that
the account listing 403 is by design, and that swift.rs falls back to the
per-account container. The protocol grid never followed. It still wrapped
the Blomp entry in `import.meta.env.DEV`, under a comment claiming Blomp is
"hidden in production until storage proxy 403 is resolved (stable: false in
registry)", which the registry has contradicted since that release.

The result was a provider shipped as production in one surface and dev-only
in another. Removes the DEV gate and replaces the stale comment with what
the registry actually records.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@snyk-io

snyk-io Bot commented Aug 9, 2026

Copy link
Copy Markdown

Snyk checks have passed. No issues have been found so far.

Status Scan Engine Critical High Medium Low Total (0)
Open Source Security 0 0 0 0 0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.

@coderabbitai

coderabbitai Bot commented Aug 9, 2026

Copy link
Copy Markdown

Review Change Stack

📝 Walkthrough

Walkthrough

Blomp is now visible in production protocol selection. Oracle Cloud is now marked as stable in the provider registry.

Changes

Provider availability

Layer / File(s) Summary
Enable Blomp production visibility
src/components/ProtocolSelector.tsx
Blomp is included unconditionally in getProtocols.
Mark Oracle Cloud stable
src/providers/registry.ts
Oracle Cloud’s stable flag changes from false to true.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Possibly related PRs

  • axpdev-lab/aeroftp#496: Both changes promote Blomp to production visibility in different catalog components.
🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name Status Explanation Resolution
Title check ⚠️ Warning The title incorrectly states that Nextcloud is promoted to stable, but the changes promote only Oracle Cloud and show Blomp in the production grid. Remove “and Nextcloud” from the title, or update the changes if Nextcloud promotion is intended.
✅ Passed checks (4 passed)
Check name Status Explanation
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check ✅ Passed Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check ✅ Passed Check skipped because no linked issues were found for this pull request.
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
✨ Finishing Touches
📝 Generate docstrings
  • Create stacked PR
  • Commit on current branch
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch fix/oracle-cloud-promote-stable

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

Nextcloud carried `stable: false, // Not tested yet` since the entry was
written, so Quick Connect warned users that Nextcloud "is not stable yet and
the service may be unavailable" while ROADMAP.md lists it among the
supported presets. The two statements could not both be true.

Verified against a self-hosted Nextcloud behind nginx: connect in 934ms with
the DAV endpoint auto-detected from the bare host, quota reported through
the OCS API (unmetered total), list, mkdir, stat, 64 KB upload with a
matching remote sha256, byte-identical download roundtrip, rename,
server-side copy, 32 MB upload with a matching sha256, public share link
creation, and recursive delete. The share link returned 404 after the
delete, so nothing was left behind. Listing was cross-checked against a
second, independent Nextcloud (Tab.digital) to separate server quirks from
client behaviour.

DigitalOcean Spaces and S3Drive keep the flag: they are untested, not known
broken, and promoting them without a live run would repeat the mistake this
series fixes.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@axpnet axpnet changed the title fix(providers): promote Oracle Cloud to stable, and show Blomp in the production grid fix(providers): promote Oracle Cloud and Nextcloud to stable, show Blomp in the production grid Aug 9, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant