[ 386.7 alpha Build(s) ] Testing available build(s)

[EdoAsus]

Why do you say that?

We're talking about software development here. Bugs happen, and they require time to be resolved.

Sure, I can wait, I understand that supporting 40 models is too much, on that I meant that they do not give you an easy life, I was used to new versions coming out more quickly, honestly in the past I was more enthusiastic about the new features, now a little less, also due to some new bugs which you do not like to encounter or expect. (RT-AX88U)

 

[John Adler]

dirty flash over the latest stable ver.
after a new start everything fine so far...
no error in the log
thxs Merlin!

 

[banger]

Dirty flash to my router and node. Node seemed to reboot or wireless went down after about 10 mins after flash but seems stable after an hour (router) and 50 mins (node). Noticed the wireless driver has been updated again to May 7.

 

[Tech9]

Noticed the wireless driver has been updated again to May 7.

Any performance differences?

 

[banger]

Any performance differences?

About the same. I was mainly interested in stability on the mesh node which seemed to be about 7 days uptime on previous Feb driver.

 

[Deleted member 22229]

Any ETA for other models i got 2 that need updating ? I know be patient right.

 

[Akore]

Updated RT-AC66U B1 main router with two USB attached drives (dirty flash with drives mounted), and two Nodes... no issues... GUI seems a bit more snappy.

 

[jsbeddow]

Updated RT-AC66U B1 main router with two USB attached drives (dirty flash with drives mounted), and two Nodes... no issues... GUI seems a bit more snappy.

The "snappy GUI" is a well observed placebo effect from a fresh reboot, many users report it after each new release.

 

[Deleted member 22229]

The "snappy GUI" is a well observed placebo effect from a fresh reboot, many users report it after each new release.

There are many placebo effects when updating code. Its not a bad thing.

 

[Clark Griswald]

I'm addicted to placebos. I would quit, but it wouldn't do anything.
-Steven wright

 

[jsbeddow]

There are many placebo effects when updating code. Its not a bad thing.

Agreed, just wanted to point it out.

I'm sure most of us here feel that installing the latest point release will result in some sort of improvement, even if that improvement is sometimes hard to quantify. Warm fuzzy feeling aside...

Edit: I do acknowledge and recognize the importance of the security updates rolled into each new release, and do personally install every new version. I'm just pointing a little fun at the sometimes perceived "performance" benefits of EVERY new release.

 

[banger]

Perceived or not node wireless seems stable but only run for 2d 2h.

 

[SomeWhereOverTheRainBow]

Some models aren't ready yet for a GPL release, so I asked them to at least provide me for now for the RT-AC68U GPL so I could begin work on merging the open-sourced parts while they work on issues affecting some of the other models.

This is why I don't support 40 different models. The more models I support, the harder it is to get them all aligned. So I limit support to a few select models of my choice.

No, I have no ETA as to when they will have debugged the affected models.

That is to be expected. I think you are lucky just to be able support the models you are able to. With that same note, you are lucky just to have access to gpls early when you do.

I already readded it when I merged the GPL (and I also skipped all the others "security hardening changes" that I felt made no sense for this project), however there was a typo in the install recipe that prevented the applet from being symlinked. Already fixed locally, but I haven't tested the fix yet so it's not on Github yet.

So it sounds like they are battening down the ladies room again?

I imagine when waiting on gpls it just takes patience. @RMerlin
Thank you
For everything!

 

[visortgw]

@RMerlin posted an alpha1 image for the XT12 a couple of hours ago (as indicated in edit to first post of this thread by OP). Changelog updated to include:

"- UPDATED: Merged with 386_48966 GPL."​

 

[banger]

Well my 68 node wireless managed to stay up for 6d before curiously disconnecting. Usually it is the Binding Update in the log that does it but that seems fixed now but nothing in the log around the disconnect (5ghz) time other than a deauth due to station leaving. :|

 

[ColinTaylor]

@RMerlin Could you take a look at this long standing bug that's been reported here and here.

The issue is specifically with NAT lookback for UDP when CTF is enabled. CTF doesn't work with UDP so there's an iptables rule that's meant to disable CTF for UDP traffic. Unfortunately the rule is placed too late in the flow (mangle/FORWARD) so never gets executed. John fixed it in his firmware by moving the rule to mangle/PREROUTING.

https://github.com/RMerl/asuswrt-merlin.ng/blob/7ba946f4a66a91d9e6b2b4b3531a32b26444a97e/release/src/router/rc/firewall.c#L6357

https://github.com/RMerl/asuswrt-merlin.ng/blob/7ba946f4a66a91d9e6b2b4b3531a32b26444a97e/release/src/router/rc/firewall.c#L6632

 

[Deleted member 22229]

Merlin has added test builds for the AX56U and AX58U. Link is in the first post.

 

[JWoo]

I had compiled my own 386.6 for my RT-AX56U several months back and did a dirty upgrade to the new 386.7 alpha. So far so good. Will post if I encounter any issues.

 

[underdose]

Dirty upgrade from 386.5_2, no errors in the syslog, all devices reconnected and work as intended. Thanks @RMerlin!

 

[Deleted member 22229]

Just flashed 386.7_A1 Working excellent !! Thanks Merlin.