profile
viewpoint
If you are wondering where the data of this site comes from, please visit https://api.github.com/users/josharian/events. GitMemory does not store any data, but only uses NGINX to cache data for a period of time. The idea behind GitMemory is simply to give users a better reading experience.

dvyukov/go-fuzz 4060

Randomized testing for Go

google/starlark-go 1316

Starlark in Go: the Starlark configuration language, implemented in Go

inetaf/netaddr 533

Network address types

dvyukov/go-fuzz-corpus 115

Corpus for github.com/dvyukov/go-fuzz examples

go4org/mem 81

cheap Go type to hold & operate on either a read-only []byte or string

josharian/compilecmp 23

automate Go compiler comparisons

josharian/2015-oscon-go-perf-tutorial 20

Slides and code for Go performance tutorial given at 2015 OSCON

push eventtailscale/tailscale

Aaditya Chaudhary

commit sha 4596beb9e03e35173c5675e1a58e022965a76a95

Fix usage in speedtest command added double hyphens to the usage string. Signed-off-by: Aaditya Chaudhary <32117362+AadityaChaudhary@users.noreply.github.com>

view details

push time in 6 minutes

push eventtailscale/tailscale

Aaditya Chaudhary

commit sha e41fb60a82627f5826be47120d06c0e6396c602d

Fix usage in speedtest command added double hyphens to the usage string. Signed-off-by: Aaditya Chaudhary <32117362+AadityaChaudhary@users.noreply.github.com>

view details

push time in 16 minutes

push eventtailscale/tailscale

Simeng He

commit sha 6c39257c952bbd6547b05d104b6d5ad1d9d1a12c

More logging to see whats wrong on githubactions Signed-off-by: Simeng He <simeng@tailscale.com>

view details

push time in 16 minutes

push eventtailscale/tailscale

Simeng He

commit sha 3568b772c92b87ce25055a0a933e693ac6c2d742

reordered node killing Signed-off-by: Simeng He <simeng@tailscale.com>

view details

push time in 23 minutes

issue closedtailscale/tailscale

Change "LogIn" option in UI disabled after successful login.

A little confusing is "Log in" is active even if already logged in.

image

When we click Log in (when already logged in) but do not complete the authentication, the client cannot work well. To resolve, customers have to log out and logged back in after that connection is restored.

<img src="https://frontapp.com/assets/img/favicons/favicon-32x32.png" height="16" width="16" alt="Front logo" /> Front conversations

closed time in an hour

darshinimashar

issue commenttailscale/tailscale

Change "LogIn" option in UI disabled after successful login.

This has been replaced with "Log in as a different user" which is more clear.

darshinimashar

comment created time in an hour

issue openedtailscale/tailscale

GRE tuneling protocol support

I made PR to support GRE tunneling protocol over tailscale. This supports L2/L3 tunneling over IPv4 (GRE over IPv6 is nor supported because tailscale MTU is 1280 which is minimum value for IPv6).

https://github.com/tailscale/tailscale/pull/2162

created time in an hour

push eventtailscale/tailscale

Aaditya Chaudhary

commit sha edf81991088506e904fd964b48350e97e534f87f

Cleaning up code Renaming variables to better explain what they do, like the cli option 'time' became 'seconds'. Changed TestConfig's Time variable to TestDuration and made it a type time.Duration. Changed the display function for the Result struct to display Mbps instead of mb/s. Signed-off-by: Aaditya Chaudhary <32117362+AadityaChaudhary@users.noreply.github.com>

view details

push time in an hour

PR opened tailscale/tailscale

Support GRE IP protocol.

This PR supports GRE tunneling protocol support.

+183 -0

0 comment

8 changed files

pr created time in an hour

issue openedtailscale/tailscale

When exit node is in use, always forward DNS through the exit node's DNS resolver

We've been talking about this for a while, but I don't see a bug for it.

When activating an exit node, we currently don't change the client device's DNS settings. If you're lucky, the client's DNS server is set to something like 8.8.8.8, which will be accessible over the exit node, and it'll be fine. But if you're using a LAN DNS server (192.168.0.1), or your ISP's DHCP-provided DNS server, it might not be accessible at all.

The new "allow LAN access" feature might make a LAN-based DNS server work, but at severe cost to privacy/security. People don't expect their DNS requests to originate from the LAN's DNS server when they have an exit node enabled.

The correct fix for this is to send Internet-facing DNS over the same resolver settings that the Exit Node itself would be using.

A separate reason to use the exit node's DNS is to ensure that DNS-based load balancing works as expected.

cc: @maisem @danderson

created time in an hour

issue openedtailscale/tailscale

iOS: still displays the list of nodes when in Stopped state

Steps:

  • Install iOS app
  • Log into a busy network
  • Move the slider to Stopped state

Expected: UI clears out everything but an option to logout/login and the start/stop slider

Actual: UI still shows all the nodes that were in your network when you "stopped."

It's easy to understand how we got here, since we already have a list of nodes that we can still display. But on macOS and Windows, the node list disappears when stopped, so we should do the same on iOS.

created time in 2 hours

issue commenttailscale/tailscale

MacOS built from HEAD split DNS not working

I'm running into this exact issue as well. Also on iOS. On Linux and Windows (both direct and via WSL2 w/ tailscaled running) it works just fine.

DentonGentry

comment created time in 2 hours

push eventtailscale/tailscale

Brad Fitzpatrick

commit sha 0debb99f0852e8fa53ab6c725966f8dd99b3fbc9

tailcfg: add DNSConfig.ExtraRecords Updates #1748 Updates #1235 Updates #2055 Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>

view details

Brad Fitzpatrick

commit sha b461ba95541100a182954694b8e02a95b7c47e49

control/controlclient: fix typo/braino in error message Thanks to @normanr for noticing. Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>

view details

David Anderson

commit sha 0022c3d2e2e8dd4fd5788158a956bdbe2ecc29d4

tsweb: replace NewMux with a more flexible DebugHandler. Signed-off-by: David Anderson <danderson@tailscale.com>

view details

Simeng He

commit sha d985ceb27de521c4cb4d86ba13d79d3b683735c7

increased tests

view details

Simeng He

commit sha 482c30ec13e0413f8f729ddac5db78dacbb537b9

tstest/integration: add TCP listener two node test Signed-off-by: Simeng He <simeng@tailscale.com>

view details

Simeng He

commit sha ff63e948dcd081e33db1a46cdef345fe8fc12e55

removed comments Signed-off-by: Simeng He <simeng@tailscale.com>

view details

Simeng He

commit sha 9d61f67e2ebf4d805f5d4d05e39f28a26b9fff1a

current work Signed-off-by: Simeng He <simeng@tailscale.com>

view details

Simeng He

commit sha 492f9499d021aa669d5a443cd66b8fd52bd2f41c

retry with node startups Signed-off-by: Simeng He <simeng@tailscale.com>

view details

push time in 2 hours

issue closedtailscale/tailscale

Mac machine fail to connect other nodes and use exit node too

do an upgrade of tailscale for 3 client; Mac(1.8.5)/android(1.8.6)/pi(1.8.7);

android could connect to pi and use it as exit node successfully;

but Mac couldn't. ip for Mac(10.15.7):

100.87.176.84

Thanks.

closed time in 3 hours

Kinghack

issue commenttailscale/tailscale

Mac machine fail to connect other nodes and use exit node too

yea. got it. just it works well before version upgrade actually. anyway, could close issue first. thanks for responding quickly.

Kinghack

comment created time in 3 hours

PR opened tailscale/tailscale

integration vm tests: run on every commit to main

This is an experiment to see how often this test would fail if we run it on every commit. This depends on #2145 to fix a flaky part of the test.

Signed-off-by: Christine Dodrill xe@tailscale.com

+17 -6

0 comment

1 changed file

pr created time in 3 hours

create barnchtailscale/tailscale

branch : Xe/vm-test-on-main

created branch time in 3 hours

issue commenttailscale/tailscale

Mac machine fail to connect other nodes and use exit node too

The exit node feature as currently defined is intended for cases where one is out in public, perhaps using a public Wi-Fi hotspot, where that Wi-Fi network is not really trusted. Once enabled, all traffic goes to the exit node and access to the local LAN is blocked. If the DNS server is on the local LAN, it would no longer be accessible. We're actually about to change this in the 1.10 release, making local LAN access be optionally available with a selection in the menu bar.

In the current release, you'll need to use a global nameserver like 8.8.8.8 or 1.1.1.1. Most likely, Android is set to use a global nameserver (there are several features in Android which can cause it to switch to global nameservers).

Kinghack

comment created time in 4 hours

push eventtailscale/tailscale

Simeng He

commit sha e5aff637a5e55e3d298d8f1402d03033f7a439b5

current work Signed-off-by: Simeng He <simeng@tailscale.com>

view details

push time in 4 hours

CommitCommentEvent

Pull request review commenttailscale/tailscale

derp: rate limit how often same-key clients can kick each other off server

 func (s *Server) accept(nc Conn, brw *bufio.ReadWriter, remoteAddr string, connN 		c.info = *clientInfo 	} -	s.registerClient(c)+	for {+		ok, d := s.registerClient(c)+		if ok {+			break+		}+		s.clientsReplaceSleeping.Add(1)+		time.Sleep(d)

should it sleep d + (something small) or is the rate limiter immune to scheduler quantization type issues where it wakes up only to discover it needs to sleep a bit longer?

bradfitz

comment created time in a day

issue commenttailscale/tailscale

Windows7-64 bit is failing on exitnode.

Nevermind, 1.9.177 should contain the #2133 fix.

darshinimashar

comment created time in a day

issue commenttailscale/tailscale

iOS 1.9.175 crash after reauth

This is surprising mess. I did the obvious thing and stopped the array-out-of-bounds lookup (filling in blanks when it happens), and I get:

2021-06-16 12:49:31.889982-0700 Tailscale[4247:2089854] *** Terminating app due to uncaught
exception 'NSInternalInconsistencyException', reason: 'Invalid update: invalid number of rows in
section 9. The number of rows contained in an existing section after the update (19) must be equal
to the number of rows contained in that section before the update (3), plus or minus the number of
rows inserted or deleted from that section (19 inserted, 1 deleted) and plus or minus the number of
rows moved into or out of that section (0 moved in, 0 moved out).'

Which, besides being an astonishingly detailed error message, suggests that the complex logic in handleUpdate is wrong. But that logic doesn't seem to have changed recently.

I'm looking what what could have triggered this, and haven't had any luck.

crawshaw

comment created time in a day

push eventtailscale/tailscale

Aaditya Chaudhary

commit sha 6af4a86f172c8d94204f34427dd4f6e958479857

Clean up code Removed printlns and commented out code, rewrote function comments for rewritten functions. Signed-off-by: Aaditya Chaudhary <32117362+AadityaChaudhary@users.noreply.github.com>

view details

push time in a day

push eventtailscale/tailscale

julianknodt

commit sha 70804eb2b75957255c6487507e1a1e59836fc992

Simplify This reduces the complexity of the priority queue and whatnot and just makes it a much simpler API. Signed-off-by: julianknodt <julianknodt@gmail.com>

view details

push time in a day

push eventtailscale/tailscale

Brad Fitzpatrick

commit sha cd282ec00fa4fbb071a40627c7a1f285aa1322e7

tailcfg: add DNSConfig.CertDomains Updates #1235 Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>

view details

David Anderson

commit sha ac07ff43bf534e5d2c2d3d8a2f18723a615b65ff

cmd/tailscaled: start after NetworkManager and systemd-resolved. The dependency is a "soft" ordering dependency only, meaning that tailscaled will start after those services if those services were going to be run anyway, but doesn't force either of them to run. That's why it's safe to specify this dependency unconditionally, even for systems that don't run those services. Updates #2127. Signed-off-by: David Anderson <danderson@tailscale.com>

view details

David Anderson

commit sha e7164425b316b447e0c6d4fa4a6df57404fe8369

net/dns: don't use NetworkManager for DNS on very old NetworkManagers. Fixes #1945. Signed-off-by: David Anderson <danderson@tailscale.com>

view details

David Anderson

commit sha 320cc8fa2191b1d04bcc54fd605db2d6d683405f

net/dns: verify that systemd-resolved is actually in charge. It's possible to install a configuration that passes our current checks for systemd-resolved, without actually pointing to systemd-resolved. In that case, we end up programming DNS in resolved, but that config never applies to any name resolution requests on the system. This is quite a far-out edge case, but there's a simple additional check we can do: if the header comment names systemd-resolved, there should be a single nameserver in resolv.conf pointing to 127.0.0.53. If not, the configuration should be treated as an unmanaged resolv.conf. Fixes #2136. Signed-off-by: David Anderson <danderson@tailscale.com>

view details

David Anderson

commit sha 93378260115b77ebfbed050e23aadbae2c3e0963

net/dns: fix inverted test for NetworkManager version. Signed-off-by: David Anderson <danderson@tailscale.com>

view details

David Anderson

commit sha 72343fbbec811d50b7909ab558aefe0f0b32f3e3

tsweb: register expvars once at startup. Signed-off-by: David Anderson <danderson@tailscale.com>

view details

David Anderson

commit sha 48c25fa36f25953bb978a697ebde33baf79d4519

tsweb: fold StdHandlerOpts and StdHandler200s with StdHandler. Signed-off-by: David Anderson <danderson@tailscale.com>

view details

Maisem Ali

commit sha 2919b3e3e6e64a5b36dac73edfcd791510a79b50

wf: loopback condition should use MatchTypeFlagsAllSet. Signed-off-by: Maisem Ali <maisem@tailscale.com>

view details

Maisem Ali

commit sha f482321f67f21a1d3bc07e780c9421181c3183b8

ipn/ipnlocal: support exit node local access switching on darwin. Signed-off-by: Maisem Ali <maisem@tailscale.com>

view details

Christine Dodrill

commit sha e0f0d10672db2d1a9786df294dbaac8b1c97f6e4

tstest/integration/vms: log to t.Logf directly (#2147) Previously we used t.Logf indirectly via package log. This worked, but it was not ideal for our needs. It could cause the streams of output to get crossed. This change uses a logger.FuncWriter every place log.Output was previously used, which will more correctly write log information to the right test output stream. Signed-off-by: Christine Dodrill <xe@tailscale.com>

view details

Simeng He

commit sha 5a7440208430cc4ac6e852579f6231ed036c463c

increased tests

view details

Simeng He

commit sha d5a6bbd7d1792369d3a4eaaa4ca47513edac260e

tstest/integration: add TCP listener two node test Signed-off-by: Simeng He <simeng@tailscale.com>

view details

push time in a day

create barnchtailscale/tailscale

branch : bradfitz/derp_kick_limit

created branch time in a day

push eventtailscale/tailscale

julianknodt

commit sha 53aa3a737a00724834bb6d6644a8639f1a7d02b3

Move upnp portmap to separate fn This isolates the upnp portmapping to another function Signed-off-by: julianknodt <julianknodt@gmail.com>

view details

push time in a day