[DSL-AC68U] GNUton / Merlin builds for DSL routers - v386 ALPHA testing only

[r0bbieNZ]

Here is a new pre-release
https://github.com/gnuton/asuswrt-merlin.ng/releases/tag/386.00_0-gnuton0_alpha0
- TrendMicro should work now. Version is shown
- JFFS seems to me still buggy in the WEB interface.
- Open NAT - not tested yet , it has been reported as not working in the previous pre-release
- AI Meshing -not tested yet
- DSL - works (using right now)

TrendMicro tested and is ok

Im having trouble formatting a new usb - detects fine - format using amtm (tried both ext 3/4) then after reboot says no device - can someone else confirm (might be just me doing something wrong)

 

[GNUton]

@RMerlin
how do you handle the prebuilts that do not have a model folder?
the problem is that tools/copy-prebuilt is copying some prebuilts (eg the ones below) in the same dir for many different models.
Do you recommend me to overwrite these prebuilds with the ones coming from DSL-AC68U GPL or should I keep the ones you upload to master?

both modified: release/src/router/et_arm/prebuilt/et.o
both modified: release/src/router/nas_arm/nas/prebuilt/nas
both modified: release/src/router/wps_arm/prebuilt/wps_monitor

 

[GNUton]

TrendMicro tested and is ok

Im having trouble formatting a new usb - detects fine - format using amtm (tried both ext 3/4) then after reboot says no device - can someone else confirm (might be just me doing something wrong)

was it working before? I will double check that soon.

 

[L&LD]

Does it work if you use another USB port?

Does it work if you remove it, wait a minute and reinsert it?

 

[GNUton]

New pre-release available for testing

Release Unstable: 386.00_0-gnuton0_alpha1 · gnuton/asuswrt-merlin.ng

Changes from the latest stable gnuton build (384.17_0-gnuton1): build: Added circleci config file build: updated tools/copy-prebuilt to support DSL-AC68U target build: fixed tools/copy-prebuilt sc...

github.com

-ASD Disabled => JFFS script in the web now looks right
- Merged latest upstream code
-Added a few missing prebuilts

 

[RMerlin]

@RMerlin
how do you handle the prebuilts that do not have a model folder?
the problem is that tools/copy-prebuilt is copying some prebuilts (eg the ones below) in the same dir for many different models.
Do you recommend me to overwrite these prebuilds with the ones coming from DSL-AC68U GPL or should I keep the ones you upload to master?

both modified: release/src/router/et_arm/prebuilt/et.o
both modified: release/src/router/nas_arm/nas/prebuilt/nas
both modified: release/src/router/wps_arm/prebuilt/wps_monitor

These are getting overwritten by copy-prebuilt because they are not model-specific, the same binary is used for all models that share the same SDK (6.37 in this case). For these, Asus has SDK-specific directories within router/ , such as nas_arm_7114, nas_arm_7, etc...

When all models/SDK are within the same folder (like rc, or asuswebstorage), then I create either directories or symlinks within prebuild/. Symlinks with platform-specific directories are used when it's CPU-specific but not model-specific, like asuswebstorage. I am doing it this way since 386 so I can have one uniform way of modifying the Makefiles (i.e. they all use prebuild/$(BUILD_NAME)/ now to access the correct binaries, no more juggling with $(HND)$(BCM)$(BCM7)$(WAHTEVER)).

 

[r0bbieNZ]

New pre-release available for testing

Release Unstable: 386.00_0-gnuton0_alpha1 · gnuton/asuswrt-merlin.ng

Changes from the latest stable gnuton build (384.17_0-gnuton1): build: Added circleci config file build: updated tools/copy-prebuilt to support DSL-AC68U target build: fixed tools/copy-prebuilt sc...

github.com

-ASD Disabled => JFFS script in the web now looks right
- Merged latest upstream code
-Added a few missing prebuilts

Tend Micro still working ok
USB drive - came up fine once new firmware started up

 

[kernol]

New pre-release available for testing

Release Unstable: 386.00_0-gnuton0_alpha1 · gnuton/asuswrt-merlin.ng

Changes from the latest stable gnuton build (384.17_0-gnuton1): build: Added circleci config file build: updated tools/copy-prebuilt to support DSL-AC68U target build: fixed tools/copy-prebuilt sc...

github.com

-ASD Disabled => JFFS script in the web now looks right
- Merged latest upstream code
-Added a few missing prebuilts

Hi @GNUton - great to see the version number updated to the 386 code branch - I'm sure that will spark more interest from the DSL "club" who just take a quick glance at version numbers without digging further.

I do however find your release notes confusing because they reference so many router model numbers that have zero to do with our DSL-AC68U's ??? Those notes seem more likely to include the background work that RMerlin has to do in order to release his code for the many router versions that he supports?

Are you actually re-doing all the work that he does in order to arrive at code that works for our DSL-AC68U ???
Forgive the stupid question - but as a non-coder I am intrigued as I would have thought you would just need to add the DSL portion of the code to the latest RT-AC68U code that RMerlin produces. [not that it's easy to do that - I'm sure it's not].

REALLY appreciate all that you have done and continue to do for the DSL folk - keep it up .

 

[GNUton]

Hi,
I will cleanup the commits and then the changes soon. I have already started with branches.

What do you mean with "I do however find your release notes confusing because they reference so many router model numbers that have zero to do with our DSL-AC68U's ???" Can you show me what are the commits you are talking about?
All changes i add to Merlin have now prefixed with "dsl-ac68u".
As for the release 386 I started from scratch. I took the RMerlin master branch and I am adding back the DSL 386 code to it.

As for supporting the dsl-ac68u i would actually say the work on adding the dsl part was easy and minimal. Most of the work is to deal with the changes between the rt and dsl, since they offer different features, prebuilts and they are developed by different teams in asus. So they differ in many parts other than the dsl.

 

[r0bbieNZ]

noticed my traffic is no longer reporting with the new build

 

[GNUton]

noticed my traffic is no longer reporting with the new build
View attachment 28080

IIRC it was working for me, but I will double check it.

 

[jonathan_892]

Hi Gnuton, 386 alpha1 working well. One small detail- the guest network only seems to connect if 'Access Intranet' is enabled. It seems to hang on 'obtaining IP address...' if disabled, this was for two different devices. It was not an issue on the previous alpha 3 build.

 

[kernol]

Hi,
I will cleanup the commits and then the changes soon. I have already started with branches.

What do you mean with "I do however find your release notes confusing because they reference so many router model numbers that have zero to do with our DSL-AC68U's ???" Can you show me what are the commits you are talking about?
All changes i add to Merlin have now prefixed with "dsl-ac68u".
As for the release 386 I started from scratch. I took the RMerlin master branch and I am adding back the DSL 386 code to it.

As for supporting the dsl-ac68u i would actually say the work on adding the dsl part was easy and minimal. Most of the work is to deal with the changes between the rt and dsl, since they offer different features, prebuilts and they are developed by different teams in asus. So they differ in many parts other than the dsl.

It is your release notes which I find confusing - not the commits ... see enclosed txt file copied from your release notes

Looks like a confusing amount of work done by you? in the production of the DSL-AC68U @GNUton version of the code?
These are your notes taken from your Github.

 

[GNUton]

Thanks for the feedback I will try to get better release note with one of the next releases

 

[SomeWhereOverTheRainBow]

Thanks for the feedback I will try to get better release note with one of the next releases

It looks like just a list of your commits. TBH your release notes should probably only reflect features enabled,fixed, added, or removed or if you add binary blobs. I think what users don't realize is that your list also includes merlin commits to base asuswrt-merlin, yours should also include those since yours is only a fork of his original firmware combined with your building method.

 

[kernol]

Thanks for the feedback I will try to get better release note with one of the next releases

Release notes not a biggy - best to continue to focus on the excellent stable builds you produce .

However - as suggested by @SomeWhereOverTheRainBow - you could base your's on RMerlin's for e.g.
https://www.asuswrt-merlin.net/changelog-382 - (which is for his last stable release - 384.19) and chop out the bits not relevant to your DSL version while adding the extra bits that you have fixed specifically for the DSL. .

 

[GNUton]

Release notes not a biggy - best to continue to focus on the excellent stable builds you produce .

However - as suggested by @SomeWhereOverTheRainBow - you could base your's on RMerlin's for e.g.
https://www.asuswrt-merlin.net/changelog-382 - (which is for his last stable release - 384.19) and chop out the bits not relevant to your DSL version while adding the extra bits that you have fixed specifically for the DSL. .

That require some manual work. I will do that if I can automate it.

 

[Jack Yaz]

That require some manual work. I will do that if I can automate it.

Unfortunately automation and good UX sometimes work against each other

 

[RMerlin]

Unfortunately automation and good UX sometimes work against each other

When there's only one single person working a project, automation is often the only solution to handle things you simply don`t have the time to handle yourself.

 

[Jack Yaz]

When there's only one single person working a project, automation is often the only solution to handle things you simply don`t have the time to handle yourself.

I agree - I have a fair number of scripts on the go here not to mention my projects elsewhere so my manual release notes aren't ideal. It does avoid "how do I understand commit history" questions though!