I am Roshan Salian and looking forward to contribute to SeaMonkey project. I would like to help with QA.

Is this community active?

roshan[m] yes it is active

roshan[m] what would you like to do? bug verification or testing?

Either is fine honestly

I landed here looking to contribute to open source/mozilla.

and I found SeaMonkey interesting and aligned with my skills.

roshan[m] SeaMonkey is actually not mozilla. We still share some resources like bugzilla and the source repos with Thunderbird and Mozilla but on the other side this is no longer any real interest in this. So we slowly move away with our own infrasturcture and services. If you want to help out you can look at bugzilla

Some unconfirmed bugs need confirmation but some are about website breakage and I frankly don't look at them any longer. I know that google and company break us frequently and we need to expand support for better web compatibility.

If you want to build SeaMonkey it is a bit compicated. Basically the 2.53 tree is the only one fully working. comm-central is frequently updated but we still play catch-up.

I will first try building the 2.53v.

Thanks

The 2.53.9b1 sources in our gitlab repo are the most up to date right now unless you want to go bleeding edge with a hg repo based on Gecko 56 and about 10000 patches :) Final 2.53.9 branches are already there but not populated yet. 2.53.9b1 tarball can also be used.

roshan[m] Which OS Linux or Windows?

Windows

You can check out mozilla central and use mach boostrap to get a working environment. mozbuild 3.3 is supported. We also support VS2019 for x64 and VS2017 for x86 and x64 builds. SDK up to 19041. The latest one seems to break shared mailnews code.

Preliminary support for VS2022 will be in the next release.

Some mozconfigs you can use:

Thanks, I will start setting up the environment.

mapi headers are now integrated with 2.53.9. You no longer need to download them. MS has also removed the source package. If you run into problems send me an email. I have a prelininary setup gude for compiler and tools. But a but outdated now.

I will contact if I am stuck. Thanks.

^bit outdated :)

hi WaltS48

hi WaltS48

Meeting in 30 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2021-08-22

Hi WaltS48

Hi WG9s_ and IanN

I am here for the meeting!

since my conflicting things were all cancelled due to impending hurricane!

I was just going to ask about that meeting, Didn't see the forwarded email, or am I just impatient.

WaltS48: just go to wiki.mozilla.org/SeaMonkey/StatusMeetings select the most recent meeting and it will say when next meeting is

Henri has been downgraded.

or will be the date of next meeting

I have the meeting in my calendar, but...

I don't get any email announcements

WaltS48: it appeared on dev-apps-seamonkey list

Well, it was forwarded to the a.c.s.seamonkey newsgroup.

WaltS48: i know if you miss a meeting you might miss announcement that due to vacations or local holidays the next meeting is in three weeks instead of two

WaltS48: yeah, looks like it was missed on the newsgroup this time around

WaltS48: anyway you are at the right place at the right time!

:)

both of my events were outside so rain and potential flooding was enough to cancel both things!

When I lived in New Orleans I went out on a date in a tropical storm. The lady decided the weather was too bad and wanted to go back home.

WaltS48 I forgot to repost in a.c.s.seamonkey. But the last time basically no feedback and so not important I think. Usually I see the same faces here :)

WaltS48: so a win there?

or was it she wanted to go back home alone?

We went to separate homes. Don't recall ever seeing her again, but I am old.

Meeting in 12 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2021-08-22

Back then I was younger and dumb enough to go out in weather like that being from the north.

Now, I'm back north and don't go out in a drizzle.

WaltS48 I still like to do this occasionally. Feeling a bit more alive when the world seems to end flooded around me :)

I'm going to read the notes. Just checked and I need to get ready to catch a couple bands at Picklesburgh since I rely on public transportation to get around. Have a good meeting!

Meeting in 3 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2021-08-22

Hi everyone

Meeting time - wiki.mozilla.org/SeaMonkey/StatusMeetings/2021-08-22

hi rsx11m

Hi rsx11m

hi rsx11m

hi WG9s_ frg IanN

who's taking notes?

me

thanks frg

late as usual probably

nominations for FotFT?

Rainer Bielefeld for support work

seconded

Action Items

Nothing new I think.

bau

Still have libera fixes on the radar but always sidetracked or something else coming up.

yeah

Status of the SeaMonkey Infrastructure

Updates seem to be not working currently.

i've just emailed ewong about it with the error message

ewong stated that they are a bit slow recently so I thought it was latest beta only but all over the board.

i've checked with both 2.53.7.1 and 2.53.8

2.53.6 yesterday too

looks like a server side issue though so might not be a reason to delay release

No did work with the releases previously so server side.

ewong: would be looking at the updates server as part of the release so best he has a heads up

anything on rust?

or VS?

rust 1.54 is supported with 2.53.9b1 and up. VS2022 will be supported in 2.53.10.

Status of the SeaMonkey Source Tree

VS2022 is still prerelease stuff and I am not a compiler expert but builds and runs fine. Mostly the same as VS2019 but now moved to the 64 bit program files dir.

All building I thinl. I started to sync 2.57 with the latest 2.53 fixes. Slow go because mozilla reformatted ESR60. Neither the original nor the 2.53 patches for c c++ stuff apply clean.

^I think

yeah, a painful process

frg: and some fun and games with intl?

IanN there will be. When I was younger and dumb 3 years ago I initially backported intl fixes without takes some prerequisites.

Need to fix this for further applying Spidermonkey fixes. Takes too much time.

Maybe next week.

Release Train

hi

Hi njsg

hi njsg

As long as you are not the late njsg :D

frg: c.tenor.com/If4TI1VXxzMAAAAM/smack-slap.gif

the ghost of njsg past!

so we are expecting to release 2.53.9 during the next week

target is 25th but we will see how things go

Probably finializing fixes which will go into it today. Also translators have updated some languages. I think the are all mostly complete now.

cZ still has a few rough edges in some

cZ didn't have strings exposed until recently so some locales were behind anyway

Yes and more changes to come. So better than before in any case.

yes, keep on keeping on

Extension tracking

github-wc-polyfill is on version 1.2.3

what is left on Flash removal?

Much. When backporting I don't take npapi but but need to be careful.

Even not working still want to keep eme support

ok

is eme tied to flash?

thought that was for html5 videos...

No but ppapi plugin support.

So just can't rip out everything with plugin in the name.

ah, interesting

easy to confuse npapi with ppapi

Always testing with a Firefox built from our code when I do rip outs. CDM is now at V10 but this can be done. Waterfox already did. I think some mouse support might be broken with Amazon watching free trailers at least. Pause needs keyboard interaction.

But overall works.

EME is still crap but we know how much bomus points we get by making a stand against it ...

Well not working anyway in SM currently.

IanN I thought it was obvious npapi was n for netscape aka Mozilla. ppapi was p for pepper aka Google.

Hi tomman

yo frg~

coming back from a 8-hour blackout

including no less than 4 failed startups because, well, our power grid is a meme

is today meeting day?

tomman: why yes it is!

tomman in the middle of it

at extensions

2.Next, Feature List, Planning and Roundtable

wrt npapi. Want to get rid of the gtk2 stuff next. Not sure how many prerequsiites it needs.

perhaps i should have mentioned this during source tree

i thought that was abbout m/c -central

always scheming and planning as usual bau

2.53 with current gitlab patches now builds and runs on aarch64 hardware but without gecko proviler

2.57 with current gitlab patches now builds and runs on aarch64 hardware with gecko profiler

Might get an M1 mini mac too in the next 2 months. More fun to play with.

based on my success here I might try to get macOS cross compiles for macos for the new apple arm processor working

WG9s_ mac needs some more stuff first I think.

but mozilla gueus are now doing firefox for trhat processor builds on nightly

Persoanlly I also would like to not do universal binaries. And native arm might need notarization. urgkkk.

What about ARM64 on Linux?

tomman works

or should do

yes

would be a interesting experience for RPi owners

tomman: aarch64 IS arm64 so should work.

Anyone suggesting 32 bit builds? If yes I will come over and try to make some personal statements: guerilla-marketing-blog.de/images/ikea-kloppe.jpg

biggest issue our tester had was confused about autoconf and did not realize that 213 was the specific version needed so had a later version and thought that was fine

:)

2.13

all sorted now?

I have an old 8-bit processor

Have the needed patches for inlining 2.13 in but needs more build fixes first

anything for my 386SX-40 with Windows 95? :D

rust to the resuce here. No 8 bit or 16 bit ports.

I think all, uh, desktop/laptop computers I have here can do IA32, not sure about the other computers

so back to 2.Next, Feature List, Planning and Roundtable

bau

yes, 2.Next, Feature List, Planning and Roundtable

still bau :)

On feature list frg gave me a patch to look at to get webextensions working. does not look easy, but I am willing to try!

thanks WG9s_

BAU from me too

how far are we in the road to Google WebComponents?

initial tabtracker implementation.

tomman I decided we need stylo for this and busy putting backports in 2.53.

Almost finished with 57 where this was inially working but unfortunately later rust version broke this hard so still on the way.

stylo is also the reason why 2.57 still needs rust 1.37

All crap but one application with a few devs against the world will again not bring in bonus points.

So just porting the stuff.

as time permits

AOB?

bau

frg wins an award for most "bau" in a meeting ;-)

rsx11m you see I am a boring guy and just chuck along with the tide.

8-)

or float with or on the waves?

it's all those fish

yes. Slapped time and time again.

next meeting September 5th?

okay, thanks for your time everyone

next meeting in 2 weeks time, same bat channel, same bat time

almost fall already...

rsx11m: yes

time flies :-p

that is labor day weekend here in the states

i will not be there

but perhaps that is a good thing

I'm around on either 5th or 12th

ok with me either way

5th is fine by me

let's keep it on the 5th then?

yep

i'm calling this a meeting

yepp2

WG9s_: happy BBQing

c u next time rsx11m

have fun

bye

Coming up next on SeaMonkey TV...

bish, bash, bop

Holly holiday IanN

ran out of bosh, so had to use bop

More: 66batmania.com/trivia/bat-fight-words

ZLOPP!

none of my words were in the list :)

Yes unauthorized content it seems.

2 more: fastcompany.com/3055253/every-batma…nomatopoeia-in-one-alphabetical-gif

"GIF"

it can be requested as gif, but it looks like the document only mentions image/gif on a title text

So, Danbooru freezes SM/PM when you try to look at their Ruffle-enabled Flash posts

the proposed solution by someone?

"just block the Ruffle scripts"

...yeah, that's hardly a solution

also, I don't run an script blocker

(also: if Tobin were here, he would eat that poor XP holdout alive)

tomman probably needs better wasm support. Not sure. Or is it custom elements again?

frg_Away: when I interrupt the script, I get this: pastebin.com/ngiFFz6L

I see no mentions of customElements anywhere, unlike in Ruffle's test pages

danbooru/danbooru c8d7f94 this is one of their recent Ruffle updates

tomman I see something similar with some third party seo tracking scripts. oracle.com is broken without a script blocker. Probably new google inspired code.

danbooru/danbooru 55980c6 the initial Ruffle commit

tomman: I suppose the scariest part in that pastebinned message might be "1:bignumber"

njsg: yeah, JS minimizers are a gift that keep giving on and on and on...

so easy to go read the code, isn't it?

bigint support is not yet in. Nur sure about bignumber. Probably also new.

frg_Away: I mean the locations in the code, line:column

FWIW, the only third-party tracking junk that Danbooru uses are a couple scripts from Cloudflare and "js-agent.newrelic.com"

(those two get blocked by my DNS adblocker)

amazingly they have resisted the temptation of WebComponents!

so the Ruffle bit caught me off-guard

njsg Ah possibly only a few megabytes js code :)

did the 2.57 devtools debugger have the ability to "unminify" the code? but that still requires opening the devtools debugger in that page. If the code weren't minified, I could just use the line and column numbers in any text editor.

njsg 2.57 debugger might be broken. I think they switched to e10s for it at least for chrome debugging. Unsure. Unminify it will not do I am sure.

> We're using Some Random Guy's unofficial NPM package for Ruffle, since the Ruffle project doesn't publish an official package themselves. We should build our own package.

:O

Usually copy paste into notepad++ and format.

and of course I want to download that stuff to try it locally: npmjs.com/package/ruffle-mirror

but That's Not How It's Done on NPMland

zipfiles? Nah, that's for dinosaurs

ir tells me it's about ~5MB of JavaScript (holy cow!), and 7 files

but I can't download anything of that anywhere

frg_Away: tomman: possibly customElements, indeed? github.com/ruffle-rs/ruffle/blob/HE…web/packages/core/src/source-api.ts createPlayer calls registerElement

"Install: npm i ruffle-mirror"

how about you just give me a zipfile I can unpack and test locally!?

I'm ABSOLUTELY AND POSITIVELY NOT INTERESTED into deploying a full-blown NPM setup on ANY of my machines, please

the AI detected you typed "I'm" "ABSOLUTELY" "POSITIVELY" "INTERESTED"

initializing installer...

I thought installers were for dinosaurs too...

it's made in flash

but requires realplayer to show a short animation

OK, so I have to parse JSON with my brain, no biggie

and finally I got a tarball

seven files: but the most important ones seem to be ruffle.js and a big fat ~5MB wasm blob

will try bringing up a local test then

oh...

...forgot how to embed Flash files on a webpage, because it has been THAT long...

it tries defining the custom element 999 times

oh wait, that's after it checks for window.customElements?

but I can't see the same check in danbooru.donmai.us/packs/js/130-61a5f0ed9a15966d0be6.js but perhaps it's in another file?

I managed to do a local test

I get no hang

instead

"Serious error encountered when polyfilling native Flash elements: Error: Failed to assign custom element ruffle-embed"

so indeed this thing wants customElements

OK, tried using the official WebComponents polyfill junk, and instead of Ruffle doing something, I get some vomit on my console

I guess I'm misfit for the JS ecosystem :)

of course the one and only WebComponents polyfill that seems to exist is the official one

and it seems Ruffle doesn't like it

...got it working!

don't try using Ruffle's "plug and play" legacy mode

instead just instance it through a script

it WILL take quite some time to load the SWF

npmjs.com/package/ruffle-mirror => use the JS API way

got it working with the standard Google WebComponents polyfills

for whatever reason interactivity doesn't work, tho.

I guess I should tell the Danbooru guys to just add the polyfills, just to be dismissed with the customary "update your browser!!!" BS :D

curiously enough, running the same Ruffle demo on FF78ESR complains with an "Something went wrong" box telling me that I Am Doing It Wrong™ by not using a web server because file:// has been deemed UNSAFE DANGER MINES by Google and friends

is it possible the ruffle code danbooru is using is older? I wonder if this explains why it loops instead of failing: ruffle-rs/ruffle a171883

I also imagine the minified code might do something in a different way...

the older code handled NotSupportedError by incrementing the counter and continuing?

but if .customElements is not defined, it'd still throw

except the minified code appears not to throw it

tomman: I'm leaning towards older version, the minified version has for (;;) and at some point the code was while(true)

njsg: they have updated their dependency, but that hone hasn't gone live yet on the production site

--that one

this one is probably needed to break the loop - at that time *infinite* loop

but I don't do this kind of javascript programming, so I'm probably missing some context

ruffle-rs/ruffle #3729 and as you mentioned XP: ruffle-rs/ruffle #3568

jsfiddle.net/rjmhcpkq horriblehorriblehorrible

but that's how far I got

(the rolling "circle 9" should be clickable)

oooh, so it MIGHT have happened on Pale Moon too

the test file DOES work fine on FF78ESR, including interactivity and "the SWF should render inside the specified container, and not elsewhere!"

something screwy might be going on with the polyfills, I guess

but judging from the comments on those two bug reports, their idea of "polyfill" is "fallback to the Flash Player plugin"

in that case, the problem is that ruffle getting stuck in an infinite loop was preventing any kind of fallback

it's possible there are still other issues in ruffle

At least I went from "hangs my browser" to "it can play movies, but you can't interact with them"

worthy material for a bugreport for them

so it also needs shadow DOM?

(See the last sentence in the last comment in ruffle-rs/ruffle #3729)

or is this actually considered a part of webcomponents? I need a dictionary for this :-P

it's part of WebComponents, indeed

both Custom Elements and Shadow DOM

hence why the official polyfills serve all that junk

oh, turns out that python3 -m http.server gives you an instant web server

now I can test properly on FF78ESR

and indeed the whole enchilada works fine, but then, the polyfills just get out of the way

Is that the module formerly (python2) known as SimpleHTTPServer?

yup

so even if the Danbooru folks accept bundling the polyfills, we would still end with broken Flash animations that don't really work

time to upload some screenshots

...and it had to be THAT old Flash video that made me enter into this wormhole

fitting for the mood (Javascript makes me feel like a baka)

Dumb idea: can you disable Web Components on Firefox to force the polyfills' usage?

...nope, you can't

tomman dumber idea can you get the people who posted the flash video to post in a supported format?

difficult

that Flash is at least 15 years old

and the group that did those stopped doing Flash videos somewhere in the previous decade

oh 15 year old flash is not an issue

at this stage those Flash videos reached memetic status, so...

back when flash was origianl macromedia product was a usefull way to get animated content on the web.

most likely there are transcodes floating in the YouTubes of the world

but of course, they don't have the charm of those pixel-perfect unlimited-HD vector graphics

..man, why noone ever came with a vector-based STANDARD video format for the Internet

?

...oh wait, that was Flash :/

then adobe bought macromedia and decided to monetize flash and made it evil and added it new features that added a lot of vulnerabilities never in the original product.

a vector format for videos would have been killer for animation

it would have saved considerable bandwidth, and would have infinite resolution by design

adobe decided to provide a way to inject ads into flash playback that was frougt with vulnerabilities.

was not an ad in the flash was a wah that plying bavk a flash video wout fo to an adobe server to injeft ads

and it is adobes inabilit to fix those vulnerabilities which is why borwsers no longer support flash content. but I see the point a 15 year old flash thing should be able to be viewed.

yeah, Flash was fine when people used only for cartoons, adbanners, and simple games

then Adobe came in and tried to make it into a full blown OS

complete with more bugs than the city dumpster

there was a flash viewer not a paer that only permitted seeing those similar to the JavaScript pdf viewer.

there was a flash viewer not a player that only permitted seeing those similar to the JavaScript pdf viewer.

IIRC there was Gnash, which never went anywhere - it only supported simple old animations

(Flash 7 and under, perharps?)

then Mozilla stepped in with Shumway - whatever happened to it?

but then 15 years ago woul work

oth yes sumway

and the latest rage is Ruffle, complete with the latest hipster webjunk

shumway

> If you have a question about Ruffle, you can ask for help on our Discord chat:

since when Open Source Software projects decided that the only way to ask for help is Discord?!?!?!

the newer the project, the more likely is to only offer Discord as a support channel

OK, I have a final test to do before I file a bug with Ruffle regarding the polyfills

test with... ew, IE11

nope, IE11 doesn't support ES6 arrow operator, and you can't workaround that with polyfills

so I guess it's a "old Firefox"-specifc issue?

Ah well, I'll go and file the bug with Ruffle anyway

What was the Seamonkey release that finally dropped NPAPI support?

6 or 7 I think.

2.53.7

ruffle-rs/ruffle #5118 done~

now let's pray, bleed a chicken, or something

tomman It is SeaMonkey with a capital M :)

always forget it on IRC, but I always take care elsewhere~

Hmmm, testing another SWF without interactivity renders it fine

with music et all

so I guess this could be workable, either at the polyfills side, or maybe even at Ruffle'sside

--Ruffle's side

gotta look if there are any alternative WebComponents polyfills, aside of Google's

Touhou is always a good excuse to debug issues with horrible code... NOT!

I think "horrible" might be better than "minified"

tomman: there is lightspark too, gnash and lightspark supported different versions of ActionScript, I think? Or of the AVM?

never heard of Lightspark

hmmm, they still target NPAPI

in other words: "get lost, your browser is ancient Firefox"

also, new shiny WASM stuff, so yeah, "your browser is not old, but for us it's old"

slightly less ominous answer, aka "we are all-in with this WebComponents stuff"

> This is actually quite unfortunate, as porting Ruffle to NPAPI would have actually been a good long-term option for targeting forks of old browsers that don't track web standards.

...WTF!?

at that point wouldn't rust be just another flash player?

I'd say there should at least be some interest in understanding what precisely is breaking ruffle

do they have a list of features they require?

it's possible they have a bad check or lack a check somewhere

probably not as bad as looping infinitely if customElements aren't supported, but still problematic somehow

most likely they don't even have a list of features

other than "oooh, Chrome 125 introduced $NEW_FANCY_API,gotta use it!!!"

however, I don't discard the notion that the polyfills could be broken or lacking features

I mean, they're just a stopgap measure, and it's likely they barely receive any testing these days, for the same reasons

it's possible the shortfills have some issue, it's possible there's some other feature this relies on, and it's possible their code has a problem.