Google even made their own fork of the Rust compiler and Chromium does not compile without it. The Google version of Rust is not compatible with the "actual" Rust, so I don't understand how it does not violate the license of Rust...

The Google version of Rust compiler is even worse than the official one. It basically only supports 64-bit x86 target and I think some ARM sets are also supported

Sompi: that implies Mozilla's rust gamble has failed google will just comsume any relevant rust code and make it specific to chromium

good night

HI all - here for the meeting

hi WG9s

status meeting in 22 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2024-09-08

status meeting in 10 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2024-09-08

for once, I'm early

hello .*

I actually remembered to update my preserved section ;)

hi all

greetings the frg

who is tired

everybody~!

hi all again after it is meeting time

yeah sundays are for napping but i hear there is a status meeting

status meeting time - wiki.mozilla.org/SeaMonkey/StatusMeetings/2024-09-08

I did some napping between dinner and breakfast, but might be in for some more after the meeting

I do the notes

Who's taking minutes?

thanks

me

frg: wow no one even had to ask!

Nominees for Friends of the Fish Tank

do we get any of these minutes back at any point?

If someone else wants to do do it be my guest :)

minutes are gone forever

until the next restart of the universe

k

12:01

just my $.02 but I keep saying we should make up a different channel for meetings so that the logbot can domeeting for us

also can do the meeting without interuupting new users looking for support ending up in the midst of a meeting

Nominees for Friends of the Fish Tank?

none from me this time. Hope anyone else has one

I think should be a SeaMonkey-mtg channel that is used for meetings and in the SeaMonkey topic it should say next meeting at @SeaMonkey-mtg at <time and date>

this woeked for me when I worked on MathMl

It would just be another adminstrative task I have no stomach for even if small.

Action Items

bau

Status of the SeaMonkey Infrastructure

frg: well please learn how to delegate!

does not need to be on you!

Amazon has a sale of whips now~

bau too. I think we cna decommision the CentOS 7 builder next but should probbaly do a test build on Rocky 8 next and test it a bit.

WG9s I know but it is usually you me or IanN

And with so small an audience I don't think it is worth it.

frg: yeah, should be easy to sort

tomman: ;-)

Status of the SeaMonkey Source Tree

I will try and setup a repo mirror of rocky 8 sources in case rpms start falling off the grid

Don't think so. Worst case Alma or Oracle Linux (shudder).

maybe make a kickstart specifically for an sm build env

friends don't let friends use Oracle anything~

awe man that unbreakable kernel!

Still on the Windows 2012 R2 builder. Asked ewong again about some files for the 2019 one.

Still patches with free esu patches so not urgent.

^patched

hmm my friend has a bumper sticker that says friends don't let friends drink Starbucks!

nsITobin WG9s comm central has some new patches. Does it show an ui yet?

not for WG9s yet

how is it supposed to work, starts UI automatically, needs startup address or parameters?

the real issue is my methodology was make a UI and fill in the gaps then go forward

I plan to set the 2.53 compatibility version to 128 from 91. Anyone against it? Running a build now and iother than the clownflare assholes no issues.

The UI he wupplied that confused me ws uing is new diagnostic buoy which I am still awating user documentaion on.

so I fixed instead of followed the bugs and now i have to clean that up but the patches WG9s has in are progress towards that

Persoanlly I would concentrate on 2.53. Still ltos of issues to resolve even in suite code.

I am remiss in not suppling feedback on the 4 remining of Tobin's patches that i have not landed anything on central gitlab for

Well I think I will do a simpler to add rigging up and coordinate with IanN on that devtools patch and then follow bugs my buoy thing was just an isolated entrypoint that isn't required its just nice to have

but yeah central has to take a backseat to 253

If we want to do central I think 5 more people/devs need to show up :)

I just resist going quietly ;) patches to the last!

yeah it should build and be used at a last resort if everything else goes to hell.

We hook up rust fully and mozilla is in a heap of trouble..

its the one thing UXP can never do.

Use rust in suite directly and I retire.

well stylo i mean

anyway nothing else for source from me. Still on wip and making progress but last week was reshuffle week.

Release Train

2.53.19 is out. Seems the 0.0.0.0 patch is incomplete per one newsgroup post.

issue is poeple d not get the idea of 0/0/0/0 sjupd be permitted as a source ip but not accpeted as a deitintaion.

is there a test case

So, if network.dns.blockQuad0 is true the IP addresses "0.0.0.0", "::"

and "::0.0.0.0" are blocked. Unfortunately, this list is not complete.

It is missing the IP address "::ffff:0.0.0.0" (IPv4-mapped IPv6

address). Just checked it on my Linux box and it's not blocked.

Note that "::ffff:0.0.0.0" is equivalent to "::ffff:0:0". Not sure which

one would be the appropriate representation in this case.

so its gonna be a whole blacklist?

Nothing for a .1 I think

nsITobin if you can test on Linux we can fix up the patch for 2.53.20.

Just the equivalent of 0.0.0.0

Yeah I'll get that possible again

the shole idea of 0.0.0.0 is a source ip that can be used to say use the normal to figure out how to get to the destination using the default IP routing table.

TOP-NOBUG-blockquad0-25319.patch

it was never itnended to be a valid distination IP address.

I didn't think it was

yes. Mozilla didn't patch it becuase CORS should take care of it. Should....

Personally I blame the *nix implementation :)

Extensions Tracking

bau

2.Next, Feature List, Planning and Roundtable

bau still some slow go because of personal issues.

the whole 0.0.0.0 issue is funny

except that these days even silly stuff can be weaponized

Bug 476108 has a UXP impl

tomman not if you are taken over :(

but 0.0.0.o0 should just never be allowed as a destination IP address should be a rule

I think

funny, I just hit an issue owing to CORS-related frame issues. but this is because a server is probably misconfigured...

bau too here

no problam if we protect against is, bt if the OS aoolws it then may odd things could happen

Well what extensions are considered core to SeaMonkey users aside from adblockers

hAVING NOTHING TO USE WITH THE BROWSER

oops hit caps lock

That would be where to start renewed extensions tracking

NoScript uBlock for me Import Export tools to a point and not trying to break enigmail

And the unnamed still maintained.

frg: but for enigmail could go the TB route on that front!

supporting openpgp

of what I have here probably OverbiteFF and User Agent Switcher might be considered such by some users, if the latter isn't, then maybe some equivalent

personally I'd consider Mnenhy too, although that probably needs to be checked for compatibility

WG9s yeah every second patch seems to be about it and used by 0.01% of its users. I rather keep enigmail for now :)

I will be working on a UI for SSUAOs rendering UAS irrelevant

njsg broken and it does not seem the owner is interested in fixing it up.

nsITobin: will it also allow to set the non-per-site override?

i don;t know what OverbiteFF deos but I think we need a new user agent switcher anyway

frg: which one? (I think all three have issues of some sort out of the box)

WG9s: OverbiteFF does the Gopher protocol

nsITobin buc had/has some patches but the break the current ui prefs by omitting SeaMonkey last I looked. I want to keep this just to annoy the web world :)

nsITobin: Extensions Tracking section might be a good starting point

Personally I till want ftp support but Gopher si somethiong from the last century. Sorry gopher supporters :)

frg: I like my gophers

gopher supporters can learn mq

;)

Would rather fix up Composer for sftp and other things than invest time in gopher.

So OverbiteFF is for Gopher? I will it till he gets elected to Congress!

nsITobin: question on xr, does it exclude test files?

yes

but

i could set up an alt instance with that removed

it just cut indexing time from over an hour and ahalf to 15 min on average

Extensions? Thanks for reminding me that I need to figure out a way to get the embedded jQuery working again on Redirector so I can actually configure it again...

Reminds me that I need to do a 57a1 patch for accesible. Omitted it initially and it did bite my lately. Adding it now will be bahhh...

been putting that on the backburner for so long

nsITobin: Oh but knor knowing that explains my sissued trying to use it. My builds have tests enabled. and I have not beenable to figure out many issues using yout xr

Well I can look for a way to see if I can do it by repo name

I'll figure something out so the sm tree can have tests index but not every other one

too much cpu?

nsITobin: well, as frg says, I am the only one who cares!

nah its single threaded and only will peg one core just time to index and reindex

WG9s well I care now an then but I think broken in 2.53 right now because of python issues.

no its cool i didn't even think about it

Needs a Firefox or TB build.

in UXP and before we kinda just started deleting test suites cause we had a choice to maintain the tests or maintain the code..

nsITobin: I am the only one actually trying to build comm-central and I have tests enabled so having a cross reference with tests diesabled does not help me. Of course if frg and IanN agree I can just diisable tests for my builds.

Oh wait this is for 2.53 not so much an issue there if a new central xr would need to have tests enabled.

We need them for 2.53. I only disable them because takes double the time. Not sure if we need them for central.

AOB

I'll enable test indexing on central and 253 it isn't a big deal

bau

anyeay all the 2.53 builds worked and are now archived

just sacrafice to the perl gods some offering it will be fine

awesome

so next meeting is in 3 weeks time on 29th September

thanks for your time today

cu here in between

So I will get linux building going again to test the linux side of the zero patch and then I am gonna focus the rest of the week on server operations and php development

or anything anyone needs me for

IanN: thanks

purplesyringa.moe/blog/webp-the-webpage-compression-format so, we now need to implement this crap too?

wait, that site is broken!

only the beginning of the site renders, the rest is several kilometers of blank space

webp support is mostly in

> SyntaxError: await is a reserved identifier

that guy is using webp as a format to compress HTML, then render it on a canvas or something

it's claiming huge bandwidth savings

it's also a horrible idea

> A nojs version is compiled separately and is linked via a meta refresh. Not ideal, and I could add some safeguards for people without webgl, but I'm not an idiot.

IanN_Away WG9s any other issues with the cross-reference you have noticed?

nsITobin: don't think so

No!

nsITobin: oh yes, one more thing

nsITobin: doesn't seem to list the results in alphabetical order

yeah I have never figured out why it seems to list it in REVERSE order

been that way since 2016

Hmm was that a Columbo reference?

nsITobin: so I searched for getCoordsForCellItem

"One More THing"

nsITobin: has /accessible/, browser/, /dom/, /toolkit/, /layout/, /comm/suite/ and /comm/mail/

0908|16:08:54 <+tomman> it's claiming huge bandwidth savings <-- that person is aware that HTTP can do compression, right?

nsITobin: not sure why toolkit appears before layout

njsg: yeah, but not strongly because even Brotli is not hip enough

IanN_Away: was this free-text search or identifier search

yeah I see it..

this is provided by glimpse

IanN_Away: this just seems to be how the results from glimpse are output.. if this used to be sorted in a particular way I can't determine that from the code

Doesn't mean it can't be changed or improved.. I just need to formally restart the cross-reference project

wish there were better options for hg hosting

... bitbucket had one job

nsITobin: foss.heptapod.net

IanN_Away: do you think *I* could get hosting on there.. I dunno if I am foss enough

nsITobin: just remember heptapod is not realy Mercurial. It is an hg front end with a git back end.

well I just dunno how complex it would be to setup a multi-repo hgweb instance and how heavy it would be .. it is backed with git?

well I did use git-cinnabar to import this to git so I could start the m3xr project code.binaryoutcast.com/laboratories/m3-cross-reference/src/TRUNK as my original repo was lost in 2022

damn i been working on this thing off and on for 8 years

this gonna go well connect.mozilla.org/t5/discussions/…ces-experiment-in-nightly/m-p/60519

I thought MZLA was an advertising corporation

now they pivoted from adtech into AI?

I knew this was coming

so... when it's MZLA's IPO?

for 6 months

tomman: MZLA seems to be an advertising company and a generative AI company on alternating days

bad decision doing it on my birthday

njsg: they are the every company where everyone is included unless you know jack shit about mozilla code or are named Tobin.. trust me

news of the chatbot integration reached at least me, I think, exactly the day after I learned about the "promoted content in location bar" 'feature'

there's also something about using commercial weather informaiton services, not sure if that one is on by default

Firefox: Now with AI and advertising, but we removed full extensions and full themes.

yep

The Browser War must be fought on many fronts

soon Firefox will drop the browser too

I mean, there are a few things to note here, Firefox, at least the project, isn't exactly known for avoiding breaking changes or feature removal, so moves that are less good for some users aren't new

Move! Fast! And! Break! Things!

don't forget your healthy dose of Rust®

But this isn't dropping ALSA support, this isn't dropping XUL extensions, this isn't changing the UI for the 100th time, this is adding two things that a sizable and vocal chunk of the online community which would embrace Firefox has strongly opposed

njsg: I was thinking the same thing

this is something people have been actively pointing out as evil or bad. Even if you have an idea that's actually not bad, MZLA would be expected to at least try to soften the bow

but for some reason executives in corporate environments seem, at best, to have no clue about the reputation of Generative Autocomplete, or might actually even think it's great

no idea what's going on at MZLA, what their executives think; I'd expect at least a bit of "uh... is this really a good idea?"

njsg: well /someone/ needs to strip firefox down as some sort of reference browser app so other people can build real firefox forks not just 14 different unique ways of doing "stitch-and-pile" like waterfox and zenbrowser

that is what Mozilla fears most

that said, this is the same universe where nobody at Amazon flinched when they decided to remotely remove a work from Kindle e-readers

and proceeded with the marvelous PR show that was remotely deleting nineteen eighty-four.

well through seamonkey technology we will find a portal and get back to the ACTUAL universe we all obviously belong in..

nsITobin: the problem is that there will be changes that make it so that the reference platform would have to have several levels/versions

supposed to belong in*

e.g. maybe some newer code designs are not very compatible with XUL extensions; some forks might want to focus on keeping NPAPI support

njsg: I assume you know the basic firefox feature compliement up until they started adding human lab experiments as components right?

the fact Mozilla has _executives_ tells you a lot of their priorities

that code compared to the rest of the browser app source is reletively light and would be easier to float across versions vs any modern firefox fork in existance

it is a personal mission of mine to have a basic command of every Mozilla tree from 24 to current release there is no exucse not to

so far my gap remains 60-114

that range is in prime rage for seamonkey backports as well as gaining an idea of the progression the code has generally gone

and because Mozilla employees explicitly do not want me personally to have it.

range*

njsg: thing is a sidebar adding sidebarified generative AI entry points would be a perfect addition for a Firefox Add-ons Site