swgp-go: add at 1.10.0 by flokli · Pull Request #29889 · openwrt/packages

flokli · 2026-07-01T22:50:51Z

Compiled on x86_64-linux, successfully tested on mediatek/filogic, 25.12.5.

📦 Package Details

Maintainer: @flokli (me)

Description:
This acts as a proxy for wireguard connections, providing some levels of obfuscation on top.

🧪 Run Testing Details

OpenWrt Version: 25.12.5
OpenWrt Target/Subtarget: mediatek/filogik
OpenWrt Device: glinet_gl-mt6000

✅ Formalities

I have reviewed the CONTRIBUTING.md file for detailed contributing guidelines.

If your PR contains a patch:

It can be applied using git am
It has been refreshed to avoid offsets, fuzzes, etc., using
```
make package/<your-package>/refresh V=s
```
It is structured in a way that it is potentially upstreamable
_{(e.g., subject line, commit description, etc.)}
_{We must try to upstream patches to reduce maintenance burden.}

BKPepe · 2026-07-02T10:20:54Z

+# swgp-go
+This readme should help you configure swgp-go.
+
+## Configuration
+Edit `/etc/swgp.json` with your desired configuration (Client, Server or both).
+
+If you're configuring a server, also properly allow that port in your firewall.
+
+Note the pprof feature is disabled in this package. If you try to enable it
+regardless, `swgp-go` will show an error on startup.
+
+## Start daemon
+Once you have created your config, enable and run swgp-go using its init script:
+
+```
+/etc/init.d/swgp-go enable
+/etc/init.d/swpg-go start
+```
+
+Please refer to the [Project README](https://github.com/database64128/swgp-go) for more configuration info and examples.


Why do we need this readme? :-) How to start daemon it is quite obvious how OpenWrt works.

Upstream doesn't have a canonical config file location, and as it requires some configuration on startup I wanted to write it down somewhere. And I saw some other packages also putting readme files like this in there. Is there a better location to put it, or you want me to remove it alltogether?

flokli · 2026-07-02T22:19:38Z

Rebased and added a test.sh script.

flokli · 2026-07-06T09:03:08Z

I got a notification from CI about a failing build just now, but this seems unrelated:

  2026-07-06 08:33:33 URL:https://mirror-03.infra.openwrt.org/snapshots/targets/x86/64/openwrt-sdk-x86-64_gcc-14.4.0_musl.Linux-x86_64.tar.zst [295175287/295175287] -> "openwrt-sdk-x86-64_gcc-14.4.0_musl.Linux-x86_64.tar.zst" [3]
  f03b79d8befd19863d4f05f8451d52a607222b7583ee8d3b901c6f34190a6fe1 *openwrt-sdk-x86-64_gcc-14.4.0_musl.Linux-x86_64.tar.zst
  sha256sum: WARNING: 1 computed checksum did NOT match
  openwrt-sdk-x86-64_gcc-14.4.0_musl.Linux-x86_64.tar.zst: FAILED

@BKPepe anything left from my side here?

openwrt-ai

Automated review of the new net/swgp-go package. Main items: two documentation errors in the README that would misdirect users (wrong config path and a swpg-go typo in the start command), plus a couple of minor Makefile/init nits. CI is still pending, so no build results to comment on.

Commit checks

Reviewed 1 new commit; message ("swgp-go: add at 1.10.0") matches the diff (new package addition), Signed-off-by present. No issues found.

Generated by Claude Code

openwrt-ai · 2026-07-09T05:51:11Z

+
+start_service() {
+  procd_open_instance
+  procd_set_param command $SWGP_GO_BIN -confPath $SWGP_GO_CONF -logLevel DEBUG


The log level is hard-coded to DEBUG, which is very verbose for a proxy that continuously handles traffic and will flood the system log by default. Was that intentional, or should this be a saner default such as info (optionally made configurable via UCI)?

Generated by Claude Code

It doesn't log for each packet, but logs two more lines once a new session is established. There isn't any uci config integration so far, and logging for sessions is also something dropbear does, so I'd say it's ok to keep it that way for now.

openwrt-ai

Re-reviewed after the earlier feedback was addressed (the commit was amended). Verified the fixes against the current tree:

README config path is now /etc/swgp-go.json, consistent with the Makefile conffiles/install rule and SWGP_GO_CONF in the init script.
README start-command typo swpg-go → swgp-go corrected.
Makefile TITLE:=/URL:= leading space removed.
Empty $(1)/etc/config dropped from the INSTALL_DIR line.

The -logLevel DEBUG question was answered (logging only occurs on session establishment, comparable to dropbear) — no objection to keeping it.

Reviewed 1 amended commit ("swgp-go: add at 1.10.0"); message still matches the diff and Signed-off-by is present. No new issues found.

Generated by Claude Code

flokli · 2026-07-09T21:00:03Z

Erf, seems there's a \t between the name and the version, not a space. Replaced with two separate invocations, as seen in multimedia/imagemagick/test.sh.

openwrt-ai

Re-reviewed the one new change since the last review (the amended commit updated only net/swgp-go/test.sh). All previously-raised items were addressed. One minor note on the test script's exit-status handling inline. The commit message ("swgp-go: add at 1.10.0") still matches the diff and Signed-off-by is present. CI is still pending, so no build results to comment on.

Generated by Claude Code

Compiled on x86_64-linux, successfully tested on mediatek/filogic, 25.12.5. Signed-off-by: Florian Klink <flokli@flokli.de>

openwrt-ai

Re-reviewed the one new change since the last review: the amended commit updated only net/swgp-go/test.sh. My earlier exit-status note is addressed — the version check ($2) now gates the exit via an explicit exit 1 on mismatch, and the package-name ($1) check was intentionally dropped.

Commit checks

Reviewed 1 amended commit (swgp-go: add at 1.10.0); message still matches the diff (new package addition), Signed-off-by present.
No new issues. The [ $? == 0 ] construct is fine here: these test scripts run under busybox ash (per the shellcheck shell=busybox convention in CONTRIBUTING.md), whose test/[ accepts ==, and this mirrors the existing in-tree net/tcpreplay/test.sh.

CI is still pending, so no build results to comment on.

Generated by Claude Code

openwrt Bot added Add package not following guidelines labels Jul 1, 2026

flokli force-pushed the swgp-go branch from 30456cd to 86c57a0 Compare July 1, 2026 22:52

openwrt Bot removed the not following guidelines label Jul 1, 2026

BKPepe added the AI It looks like this is issue or pull request was created by AI or it includes AI comments label Jul 2, 2026

BKPepe requested changes Jul 2, 2026

View reviewed changes

flokli force-pushed the swgp-go branch from 86c57a0 to cf1516f Compare July 2, 2026 10:42

flokli requested a review from BKPepe July 2, 2026 10:42

flokli force-pushed the swgp-go branch 2 times, most recently from 40b7487 to a7179b8 Compare July 2, 2026 22:18

openwrt-ai reviewed Jul 9, 2026

View reviewed changes

flokli force-pushed the swgp-go branch 2 times, most recently from 50cadf8 to 2a96be0 Compare July 9, 2026 07:54

openwrt-ai reviewed Jul 9, 2026

View reviewed changes

flokli force-pushed the swgp-go branch 2 times, most recently from c7ef391 to d9400d2 Compare July 9, 2026 21:00

openwrt-ai reviewed Jul 10, 2026

View reviewed changes

Comment thread net/swgp-go/test.sh Outdated

swgp-go: add at 1.10.0

859659f

Compiled on x86_64-linux, successfully tested on mediatek/filogic, 25.12.5. Signed-off-by: Florian Klink <flokli@flokli.de>

flokli force-pushed the swgp-go branch from d9400d2 to 859659f Compare July 10, 2026 07:11

openwrt-ai reviewed Jul 10, 2026

View reviewed changes

Uh oh!

Conversation

flokli commented Jul 1, 2026

📦 Package Details

🧪 Run Testing Details

✅ Formalities

If your PR contains a patch:

Uh oh!

Uh oh!

BKPepe Jul 2, 2026

Choose a reason for hiding this comment

Uh oh!

flokli Jul 2, 2026

Choose a reason for hiding this comment

Uh oh!

flokli commented Jul 2, 2026

Uh oh!

flokli commented Jul 6, 2026

Uh oh!

openwrt-ai left a comment

Choose a reason for hiding this comment

Commit checks

Uh oh!

Uh oh!

Uh oh!

openwrt-ai Jul 9, 2026

Choose a reason for hiding this comment

Uh oh!

flokli Jul 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

openwrt-ai left a comment

Choose a reason for hiding this comment

Uh oh!

flokli commented Jul 9, 2026

Uh oh!

openwrt-ai left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

openwrt-ai left a comment

Choose a reason for hiding this comment

Commit checks

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants