amtm amtm - the Asuswrt-Merlin Terminal Menu

[smasher]

No way amtm does any of that. Period.

I tried manually filling in the fields; it didn't work. Then I tried "Quick Internet Setup" from the GUI, and it still didn't work. The PPPoE login credentials kept going back to the admin login credentials. I even re-installed the settings while amtm was installed, and still, every time I filled in the PPPoE login credentials in the GUI and hit "Apply", they'd go straight back to the router's admin login credentials and fail to connect to the ISP.

As soon as I uninstalled amtm and restored by settings, everything was back to normal and working fine.

 

[thelonelycoder]

I tried manually filling in the fields; it didn't work. Then I tried "Quick Internet Setup" from the GUI, and it still didn't work. The PPPoE login credentials kept going back to the admin login credentials. I even re-installed the settings while amtm was installed, and still, every time I filled in the PPPoE login credentials in the GUI and hit "Apply", they'd go straight back to the router's admin login credentials and fail to connect to the ISP.

As soon as I uninstalled amtm and restored by settings, everything was back to normal and working fine.

Since you can read code easily, you'll notice that amtm does not start by itself unless invoked trough an ssh command. Also, there is no code in amtm that reads or modifies any of your mentioned nvram variables.

 

[dave14305]

I tried manually filling in the fields; it didn't work

Did you disable any browser auto-fill or password manager extension? There are often hidden fields that get overwritten by password managers.

I exclude my router url from LastPass fill in.

 

[L&LD]

@smasher, try logging in with 'private browsing mode' enabled. That should stop the auto-filling of fields. If you've kept that option enabled and enforced, of course.

 

[smasher]

Since you can read code easily, you'll notice that amtm does not start by itself unless invoked trough an ssh command. Also, there is no code in amtm that reads or modifies any of your mentioned nvram variables.

Yeah, installed and uninstalled via ssh/CLI.

@smasher, try logging in with 'private browsing mode' enabled. That should stop the auto-filling of fields. If you've kept that option enabled and enforced, of course.

Did you disable any browser auto-fill or password manager extension? There are often hidden fields that get overwritten by password managers.

That does sound like a good idea, to isolate if the browser is messing with it. I'll try to make time for testing it in "private mode", in the near future. Weird, though, that the browser would only mess it up when amtm is installed... But if I test it, I'll report back here.

 

[dave14305]

Weird, though, that the browser would only mess it up when amtm is installed...

Not weird, but downright impossible.

It’s likely coincidence with another reason you may have had to hit the Apply button on the WAN page. Restoring your settings likely fixed it, regardless of uninstalling amtm.

But that’s just speculation on my part since I wasn’t there.

 

[smasher]

Not weird, but downright impossible.

Doubly impossible, since I installed it twice, and uninstalled it twice, with the same results both times.

 

[ColinTaylor]

Doubly impossible, since I installed it twice, and uninstalled it twice, with the same results both times.

As all the amtm "installation" does is download the script and copy it to /jffs/scripts the only thing I can think of is perhaps your /jffs partition is being filled up leading to spurious behaviour.

 

[Adamm]

Doubly impossible, since I installed it twice, and uninstalled it twice, with the same results both times.

Unlike humans, code doesn't make mistakes or assumptions. It does what it was designed to-do. In this case what you are describing is in-fact impossible and almost certainly a browser autofill issue.

skynet@RT-AX88U-DC28:/tmp/home/root# cat /jffs/scripts/amtm | grep "nvram set"
                            nvram set jffs2_scripts=1
                                    nvram set usb_idle_timeout=0
                    1)        nvram set reboot_schedule=00000000000
                            nvram set reboot_schedule_enable=0
                nvram set apps_mounted_path=
                nvram set apps_dev=
                nvram set apps_state_autorun=
                nvram set apps_state_enable=
                nvram set apps_state_install=
                nvram set apps_state_switch=

As you can see, none of the nvram calls modify anything to-do with the WebUI or more specifically login details. Best not to spread false information/accusations so less tech savvy users may accidentally believe this.

 

[smasher]

As you can see, none of the nvram calls modify anything to-do with the WebUI or more specifically login details. Best not to spread false information/accusations so less tech savvy users may accidentally believe this.

I'm sharing my experience with it. I'm not a n00b, and the browser may be the culprit here. Hopefully I'll test that out in the near future by isolating that variable via "private browser" session.

In any case, something is screwy, and the thing to do now is find out what it is so it can be fixed and/or documented as a "gotcha".

 

[Ro berto]

Hopefully I'll test that out in the near future by isolating that variable via "private browser" session.

you can also try using a browser you have not used before (i.e. internet explorer).

 

[5stringdeath]

you can also try using a browser you have not used before (i.e. internet explorer).

A browser he's never used before, or a browser nobody uses anymore?

 

[martinr]

A browser he's never used before, or a browser nobody uses anymore?

Welcome to the forum!

 

[smasher]

A browser he's never used before, or a browser nobody uses anymore?

I'll try it with "ELinks"

 

[5stringdeath]

Welcome to the forum!

Long time lurker, first time caller.

 

[martinr]

Long time lurker, first time caller.

Quality over quantity. Look forwards to your future posts.

 

[smasher]

Did you disable any browser auto-fill or password manager extension? There are often hidden fields that get overwritten by password managers.

@smasher, try logging in with 'private browsing mode' enabled. That should stop the auto-filling of fields. If you've kept that option enabled and enforced, of course.

This does seem to be the problem.

Now it makes sense that the problem seems to have started after installing dnscrypt, then changing some settings on the GUI's WAN page, which messed up the PPPoE login credentials, and replaced them with the router's login credentials.

Looks like a "gotcha", not a bug. Still deserving of attention, or mention.

I am curious if it's possible to not let the browser confuse those different logins, eg using different field names? I don't know...?

 

[L&LD]

@smasher, you would need to ask the authors of the browsers you used.

 

[FalconB]

@thelonelycoder : I found a user interface-issue that I think can be improved. If I do a search for updates through the "u"-option and there are no updates available (ie all installed scripts and amtm itself are up to date), then amtm display the script names followed by "no upd" and the text "All scripts are up to date". Below that is the text "Enter option". This text is the same as when you first start amtm and the main menu is shown. However, at this point the user is not shown the main menu, only the input prompt, and if you dont remember the main menu commands you need to exit the script and then start it again to get the main menu to display (or do trial'n'error to enter a submenu and then go back). I suggest that after the update check, the user is prompted with a "Press enter for main menu" or similar that brings up the main menu again. Just my thougths, up to you if you want to incorporate it or not (maybe there is a reason for why it is the way it is that I don't know of). Thanks for a great script!

 

[smasher]

@smasher, you would need to ask the authors of the browsers you used.

I won't hold by breath waiting for the browser developers to fix a bug (or is it a feature?) that may only affect edge cases like this. It seems the better option, if it's feasible, would be for a fix/workaround that assumes the browser will screw it up. That's why I mentioned different field-names in the HTML forms as a possible solution.