This must be the future of JavaScript: emojicode.org

(From news://news.eternal-september.org/87ikfb9kqx.fsf⊙ti )

> Emojicode is an open-source, full-blown programming language consisting of emojis.

we're doomed

I think emojicode was first used in Idiocracy so should work :)

Status meeting in 30 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2025-11-16

Status meeting in 20 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2025-11-16

Status meeting in 10 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2025-11-16

hi .*

hi njsg

hi IanN frg njsg

rsx11m hi

rsx11m: hi

hi rsx11m

Status meeting time - wiki.mozilla.org/SeaMonkey/StatusMeetings/2025-11-16

we need a hi-bot ;-)

Who's taking minutes?

me

thanks

Nominees for Friends of the Fish Tank

njsg for finding the recraptcha bug to backport

seconded

thirded

Action Items

bau

Status of the SeaMonkey Infrastructure

bau for me.

website was switched across to the new infra this weekend, thanks to ewong for helping with that

Looked into VS2026 but it dropped generating code for 7 and 8.1 so nothing we want in the forseeable future. Same for the new runtime. But more AI crap added...

No Clippy yet?

Co Pilot is the new clippy

but more annoying than the old clippy :-(

still a shame they haven't used MS Agent as an interface for it.

mozilla is doing what for Windows releases, cross-compilation or VS?

after esr 115 it dropped 7 and 8.1 support.

But clang only for some time. VS20xx is still needed for support programes.

Cross compilation under Linux uses wine. A total mess I don't want. Unlike macOS which is smooth.

hopefully the clang part means they'll not grab something from VS that can't be used for 7 or 8.1; but yes, using wine for that sounds less than ideal

They already did with rust.

and sticking with Rocky/Redhat Linux 8.10 as later versions bump min lib versions too high

Anyway both still have some support till 2026. Somewhere down the line we might need to make a decision but not now.

And 10 needs Haswell. No more Sandy and Ivy Bridge which are still more than capable for normal work.

Status of the SeaMonkey Source Tree

Fixed some regressions with IanN and I think we found the Linux crash on close solution.

No time to look if central still builds but I can still push for now.

Release Train

2.53.22 has 2 regressions and the close crash. I think we should do 2.53.23 soon and without a beta. Not that much new stuff in compared to 2.53.22

We can delay automatic updates for a week or so but current release is very stable for me.

And making recaptcha work is worth it.

I'm confused - I saw 2.53.22 released on Friday when setting up the agenda, but now that post disappeared,

Also, the main page offers me 2.53.21 instead of 2.53.22

Was 2.53.22 officially retracted then?

possibly a problem with moving to the new infra?

rsx11m: maybe the move to the new website

Oh it looks like ewong switched to heptapod which is not updated yet. I updated gitlab but waited a bit for heptapod becuase he did some branches there.

ok, so it wasn't my imagination ;-)

and no sun spots either

maybe not a bad thing after all if 2.53.22 needs regressions fixed

They are not that bad. CSS and bugsplat version. The Linux issue does not always occur.

And only during close whith no data loss.

only needs a heptapod sync then, apparently

Extensions Tracking

bau

2.Next, Feature List, Planning & Roundtable

bau?

yau!

bau

AOB

finally done with thee DST business

bau

for this year at least...

indeed

we look forward to DST fun in March

:-p

I should try to make two or three FAQ entries out of topics in the newsgroup, let me see, adding that to the wiki should be easy

(Also, wondering how easy/hard is it to put line numbering in the text editor. Somebody said Kompozer had it, so maybe it's possible to get that somehow)

njsg uses the mozilla editor component. Now aware of anything in it. Might need backports from Kompozer.

okay, I'll call this a meeting, next meeting in 3 weeks time (7th December), same bat channel, same bat time

thanks for your time everytone

how's the plan for christmas?

*thanks for your time everyone

cu

4 week interval so that next after the 7th is in January?

I reckon we keep Christmas this yeah

I reckon we keep Christmas this year

4th January I think

either is fine with me, 28th ought to work well as will any other date around that for me

i'm probably only half around on 28th

aww, I'm late :D

tomman: happy no more America DST meeting day for you!

I live at the /other/ America, we never had stinkin' DST :D

hi tomman

tomman: I still call the whole thing America, that's how I learned continents!

oh, I thought you were following US DST dates, my bad

also, that reminds me I still need to figure out how to properly create a custom locale for libICU

lovely, The Retro Web crippled even more its interface to pander to the Chrome JS Framework crowd

I find kinda ironic (and sometimes insulting) that retrocomputing sites actually require bleeding edge, er, "modern" JS just to work

ah well, at least the search forms still let me write keywords at some places, so it's not a total loss... yet

en.namu.wiki/w/Windows%209x and speaking about retrocomputing... this wiki "crashes" with the famous Gray Screen of Death™, after briefly flashing parts of an actual wiki page

...because it's not a wiki or even a website, it's an App. ("app initialization failed")

if you manage to stop the page load on time, it will render a perfectly readable wiki page

...and now I realize this wiki could be AI slop :/

Is the regression in bug bugzilla.mozilla.org/1987293 going to be adressed some time, or is more information needed?

Ektus: difficult to troubleshoot unless you have that brand of router

I do have some entries in the error console. Would those give a hint?

Ektus: they might

Timestamp: 16.11.25, 18:03:33 MEZ

Ektus: or they could be a red herring

Error: ReferenceError: setControlsEnabled is not defined

Source File: 192.168.3.1/js/login.js?1.14.0

Line: 1

so even routers require modern javascript with chromeisms now...

Timestamp: 16.11.25, 18:04:25 MEZ

Error: Error: Permission denied for <192.168.3.1> to create wrapper for object of class UnnamedClass

Source File: 192.168.3.1/js/common.js?1.14.0

Line: 1

This error is present with versions of gargoyle as of some years back. As described in the bug, it's been a regression introduced with nightly build 2025-04-08 of 2.53.21b1

Ektus: and an earlier nightly doesn't show the error?

Exactly. I've been using nightlies for quite some time and have them archived, and when I first encountered the bug I went back in time until I found the last version that was working. And also the earliest version of Gargoyle that uses an UI modern enought to trigger the error, which is 1.10.0

Sompi: OpenWRT uses some sort of JS framework these days for LuCI, although thankfully it works on SeaMonkey without issue... SO FAR

1.10.0 was intreoduced in 2017: gargoyle-router.com/phpbb/viewtopic.php?t=11271

tomman: another problem with these HTTP frameworks with routers is that it cannot be encrypted, because you cannot get a certified encryption key for a link-local IP address... so if you try to access the control panel with Edge, Chrome or even some configurations of Firefox, it refuses to load the page either because the connection is unencrypted or not the key certified by authority

and apparently now you cannot access them all with SeaMonkey either!

why don't they just make telnet-based control interfaces to them...

Thankfully OpenWRT has SSH

I know that telnet-based interfaces used to be a common thing in the past, but at some point they just stopped making them and those HTTP-based interfaces are nothing but problems

and I refuse to buy routers that cannot be reflashed for that and other many, MANY reasons

Why use SSH, if your computer is directly connected to the router anyway? No-one is going to eavesdrop that. SSH and all encryption stuff just adds more overhead and makes things harder to implement

Plain telnet is more than good enough

Your computer is connected to the router... which is also connected to the raw Internet

hence why it won't ship with Telnet enabled out of the box, or even available

Also, sometimes there is value to be able to remote into your router

Of course the router does not accept telnet connections from the outside, only from the local link

it's an extra daemon, someone would say "extra attack surface"

and these days nothing ships Telnet anymore

The router should be configurable with the simplest possible computer

...and for The Market™, apparently said "simplest possible computer" is a cellphone - you can buy routers that have neither SSH/Telnet or even a web UI

cellphone app mandatory

Google for example had routers like that (until they killed that product line, as Google does)

The simplest possible computer may be something like some 8- or 16-bit nerdy thing with very little memory

Anyway for me SSH on a LAN is fine

A homebrew PC/XT clone or something

I've used PuTTY on a 386SX-33 running Win95

Plain telnet is always the lowest common denominator

SSHing from that thing is slow, but not THAT much (especially if running a full fat OS like 95!)

Timestamp: 16.11.25, 18:04:25 MEZ

Error: ReferenceError: addLoadFunction is not defined

Source File: 192.168.3.1/hooks/login/js/050-tor.js?1.14.0

Line: 1

Timestamp: 16.11.25, 18:04:25 MEZ

Error: ReferenceError: addLoadFunction is not defined

Source File: 192.168.3.1/hooks/login/js/070-openvpn.js?1.14.0

Line: 1

Timestamp: 16.11.25, 18:04:25 MEZ

Error: ReferenceError: UI is not defined

Source File: 192.168.3.1/i18n/English-EN/strings.js?1.14.0

Line: 5

Timestamp: 16.11.25, 18:04:25 MEZ

Error: ReferenceError: UCIContainer is not defined

Source File: 192.168.3.1/login.sh

Line: 57

Timestamp: 16.11.25, 18:04:25 MEZ

Error: ReferenceError: setChildText is not defined

Source File: 192.168.3.1/js/login.js?1.14.0

Line: 1

Timestamp: 16.11.25, 18:04:25 MEZ

Error: ReferenceError: proofreadLengthRange is not defined

Source File: 192.168.3.1/login.sh

Line: 1

From my experience SSH itself is not slow at all, even on very slow computers. The problem is the implementation that doesn't exist

Ektus: is the JS readable from the router readable or has it been minified?

Ektus: smells like a script that failed to load at all

Even an Atari ST using netsurf should be able to configure a router, but it's very slow. Forget about SSH on anything slower than 68060 CPUs. It still need lots of computing power.

probably because of the very first error on that chain

sometimes a missing feature causes a lot of stuff to fail because variables and objects do not get defined

so you get a TON of errors, but only one of those is actually relevant, and it's usually the earliest one (or close to it)

That's the router I'm currently connected with. I can try to load the script, MOMPL

minified scripts are also a PITA and a half to debug

var logS=new Object;function doLogin(){var e=document.getElementById("password").value;if(0==e.length)alert(logS.passErr);else{setControlsEnabled(!1,!0,logS.Lging),sessionExpired=!1,passInvalid=!1,loggedOut=!1;var t=getParameterDefinition("password",e);runAjax("POST","/utility/get_password_cookie.sh",t,(function(e){if(4==e.readyState){if(e.responseText.match(/^invalid/))passInvalid=!0,setStatusAnd

Quotas();else{var t=e.responseText.split(/[\n\r]+/),l=0;for(l=0;l<t.length;l++){var a=t[l].replace(/^.*ookie:/,"").replace(/\";.*$/,"");a.match(/=/)&&(document.cookie=a)}window.location.href=window.location.href}setControlsEnabled(!0)}}))}}function checkKey(e){var t=0;window.event?t=window.event.keyCode:e&&(t=e.which),13==t&&doLogin()}function setStatusAndQuotas(){setChildText("current_time_date",

cnv_LocaleTime(currentTime)),sessionExpired?(setChildText("login_status",logS.SExp,"red"),login_status.style.display="",login_status.className="alert alert-danger"):passInvalid?(setChildText("login_status",logS.InvP,"red"),login_status.style.display="",login_status.className="alert alert-danger"):loggedOut?(setChildText("login_status",logS.LOut,"black"),login_status.style.display="",login_status.c

lassName="alert alert-danger"):setChildText("login_status","","black");var e,t=[],l=[],a=[];for(e=0;e<quotaIdList.length;e++){var n=quotaIdList[e],s=quotaIpLists[n];s.length>0&&("ALL"==s[0]?t.push(n):"ALL_OTHERS_COMBINED"==s[0]?a.push(n):testAddrOverlap(connectedIp,s.join(","))&&(l.push(n),localIpName=connectedIp))}var o,i=["black",!1,"11px"],d=["red",!0,"12px"],r=[l=0==l.length?a:l,t],c=["local_q

uotas","global_quotas"];for(o=0;o<r.length;o++){var p,u=r[o],m=c[o],g=0==o?connectedIp:"ALL";if(clearFieldset(m),u.length>0)for(p=0;p<u.length;p++){var h=u.length>1?p+1:-1,y=createQuotaDiv(u[p],g,h,i,d);document.getElementById(m).appendChild(y)}document.getElementById(m).style.display=u.length>0?"block":"none"}}function clearFieldset(e){var t=document.getElementById(e);if(null!=t){for(var...

...l=null;null!=t.firstChild;){var a=t.firstChild;l="panel-heading"==a.className||"panel-title"==a.className?a:l,t.removeChild(a)}null!=l&&t.appendChild(l)}}function createQuotaDiv(e,t,l,a,n){var s="ALL_OTHERS_COMBINED",o=quotaIpLists[e],i=0;for(i=0;i<o.length;i++)s=testAddrOverlap(t,o[i])?o[i]:s;var d=quotaUsed[e][s],r=quotaPercents[e][s],c=quotaLimits[e][s],p=document.createElement("div");if(p.c

lassName="panel-body",l>0){var u=document.createElement("span");u.style.fontSize=n[2],u.style.display="block",u.style.fontStyle="italic",u.style.textDecoration="underline",u.appendChild(document.createTextNode("Quota"+l+":")),p.appendChild(u);var m=timeParamsToLines(quotaTimes[e]),g=document.createElement("span"),h=document.createElement("span"),y=document.createElement("span");for(h.appendChild(d

ocument.createTextNode("Active "+("only"==quotaTimes[e][3]?"Only:":"All Times Except:"))),h.style.fontSize=a[2],h.style.marginLeft="25px",h.style.display="block",h.style.width="150px",h.style.cssFloat="left",h.style.styleFloat="left",y.appendChild(document.createTextNode(m.shift())),y.style.fontSize=a[2],y.style.marginLeft="25px",y.style.display="inline",g.appendChild(h),g.appendChild(y),p.appendC

hild(g);m.length>0;){(y=document.createElement("span")).appendChild(document.createTextNode(m.shift())),y.style.fontSize=a[2],y.style.marginLeft="200px",y.style.display="block",h.style.clear="left",h.style.clear="left",p.appendChild(y)}}var f,v=logS.Qnam;for(f=0;f<3;f++)if(c[f]>=0){var C=v[f],x=parseBytes(c[f]),S=x.replace(/^.* /,""),L=parseBytes(d[f],S).replace(/ .*$/,""),E=truncateDecimal(r[f]);

L=d[f]>c[f]?x.replace(/ .*$/,""):L;var T=document.createElement("span"),_=document.createElement("p");l>0?_.appendChild(document.createTextNode(E+"% "+logS.of+" "+C+" "+logS.fQuo+l+" "+logS.husd+" ("+L+"/"+x+")")):_.appendChild(document.createTextNode(E+"% "+logS.of+" "+C+" "+logS.qusd+" ("+L+"/"+x+")"));var I=100==r[f]?n:a;_.style.color=I[0],_.style.fontWeight=I[1]?"bold":"normal",_.style.fontSiz

e=I[2],T.appendChild(_),T.style.display="block",T.style.clear="left",l>0&&(T.style.marginLeft=l>0?"25px":"0px"),p.appendChild(T)}return p}function timeParamsToLines(e){var t=e[0],l=e[1],a=e[2],n=[];return"always"==e[3]?n.unshift(UI.Always):""!=a?n=a.match(",")?a.split(/[\t ]*,[\t ]*/):[a]:(""!=t&&(n=t.match(",")?t.split(/[\t ]*,[\t ]*/):[t]),""!=l&&n.unshift(l)),n}

Ektus: probably best loading into a pastebin

That's 192.168.3.1/js/login.js?1.14.0

The Gargoyle router software is open source. github.com/ericpaulbishop/gargoyle

1116|17:18:29 <+tomman> hence why it won't ship with Telnet enabled out of the box <--- yeah, Huawei begs to differ on that one...

1) put that on a pastebin, and 2) on a separate paste put the ENTIRE error log too - open a clean tab, fire the inspector, and load the page so only the relevant errors get logged

Ektus: oh, so this is a 3rd party firmware?

and by that I mean to the WAN

@tonman: Yes, it's openWRT based, but with more QOS in mind. It's been around for a long time.

also perhaps error log from the working version, if that has lots of errors too, in case those unique to the failing builds are asmall number

Ektus: then it must be using a different web UI layer, that's definitely not LuCI which is what OpenWRT generally ships

Of course it's not. It comes from a time when Luci was not shipped by default, even.

(I use OpenWRT 23.05 here on a Archer C7v5)

LuCI also uses a similar JS junk, and won't work at all if you disable JS :/

Also using Archer C7. And 1043ND. And C6. And some more. And almost all with Gargoyle if possible.

(DD-WRT is less reliant on JS, but it lacks support for anything Mediatek)

@tomman: There are just a couple errors, and some more warnings. brb, switching to an unaffected version to check the error console there...

Huawei's routers also don't work with NetSurf at all

tomman: come across Tailwind?

Now running an older build (Build identifier: 20250331211856). The router login is working, with zero errors and a couple warnings.

ektus.de/Gargoyle_Seamonkey_debug.zip containing bootstrap.min.css, common.css, theme.css and login.js

Info has been attached to the bug. Archive is hosted on my own webspace, don't have a pastebin account. Files should be under GPL.

Plenty of free pastebins out there, I use paste.debian.net/1

but for multiple stuff that needs to be persistent, yes, it's best to host them yourself

Is there a way to copy the errors and warnings from the Error Console in one go?

not really

now, from the dev console (F12->Console), a workaround to expand ALL records, then Select All->Copy

but from the actual Error Console window, nope, there is no Select All (you can't even multiselect there!)

this has been like that since the early Mozilla Suite days :/

That's what I found. Just 9 errors with my bug, so easy to copy individually, but lots of warnings. But those I should better report on the Gargoyle side anyway, I guess.

gn8