Meeting time in 27 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2024-09-29

Meeting time in 18 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2024-09-29

Meeting time in 4 minutes - wiki.mozilla.org/SeaMonkey/StatusMeetings/2024-09-29

hello .*

hi njsg

hi rsx11m

hi IanN njsg

Hi rsx11m

hi frg

rsx11m: hi

Meeting time - wiki.mozilla.org/SeaMonkey/StatusMeetings/2024-09-29

Who's taking minutes?

me

Thanks frg

Nominees for Friends of the Fish Tank?

rsx11m: w as wondering what happened to you glad to hear it was just a google account issue. two meetings in a row no meeting announcement from you.

yeah, Google is doing funny things occasionally ...

Myckel for foxong the linux clang later than version 15 can not find the X11 libs

seconded

thirded

thirding Myckel for FotFT ^

s@thirding@fourthing@

was an over a year issue

Action Items

bau

Status of the SeaMonkey Infrastructure

Mostly bau. ewong copied the stuff from the old Windows build server over but I didn't found the time yet to move it in.

another week or so and will start looking at switching from aus2 to aus5

Status of the SeaMonkey Source Tree

2.53 now has the same build fixes as wip. Only some js stuff like cranelift missing.

More release but when I uploaded the latest sym files to bugsplat I noticed that the Windows x86 has about 8 times the crashes compared to the x64. 90% OOM. And given that the x64 probably has a laerger user base even worse.

Persoanlly i think x86 should go soon.

how easy is it to track down OOM?

saw a test post from a "WOW64" build in eternal-september.test, I'll see if I reply to warn about that, it either is intentional or isn't :-)

but then x86 is kind of not a priority on firefox so perhaps keeping it would be a differentiator

I reprocessed the dumps and ti was pretty obvious.

Just looking at the heading

IanN: another question, without wanting to get too deep into this here is: are there options that could customize how allocation works? I think I saw some js, perhaps GC-related, entries somewhere (mozillazine?) but didn't check if these apply to the 2.53 code

Well I rather differentiate elsewhere like in OS support and no telemetry :)

so WOW64 is a very confusing thing. this has 64 in it but is for running 32 bit apps on 64 bit windows

(Alongside, disabling or limiting on-memory cache could maybe help a little. but my experience with 2 GiB on GNU/Linux is that it's quite easy to hit thrashing if SeaMonkey is running for long enough, even with NoScript. Now maybe AdBlock Plus *is* a problem...

oddly not sure this is still the case but when ehy first made this up system32 was for 64-bit apps and WOW64 was for 32 bit apps.

WG9s: yeah, this might be one good thing out of it, if it says WOW64 it tells you it's likely running an IA32 build on an amd64 system

njsg bloated sites just need memory these days. I even added some patches which might have made it worse but helped with x64 crashes.

njsg I looked at the buildid so it was a clear cut.

maybe need to check how many people are still running 32-bit on 64-bit OS

Some are even running the x86 under Windows 11.

WG9s: it may help to think of it as the full "Windows on Windows 64", but it's still far from an ideal acronym...

somehow migrate them to the 64-bit version...

should have always bee W32OW64

rather than WOW64

nd also never had 64 bit stuff in system32

the second one might have been harder to do but the first would have been dead simple

anysay back to SeaMonkey

anyway

It is not an urgent problem and I don't want to altert he source but I think these builds are obsolete. Even more when we build on Rocky 8 and raise the library versions needed.

But something to keep in mind maybe for 2025.

hi nsITobin

I'm not sure automating migration is a good idea, I was thinking of showing an alert in the installer, but not sure if that won't be considered intrusive, but then a problem is those who are using 32-bit and update it automatically... can the update system show notices of some sort?

sorry to be late

frg: how many crashes were there?

(i.e. WOW64 OOM crashes)

(or IA-32 in general)

680 versus 26 for Windows Linux x64 was low too. I don't have an exact count for OOM only. x64 had some too.

Status of the SeaMonkey Source Tree

86 not 26

I been crashing status meetings since 2008 ;)

bau. Raised the ua minimum version to 128 for 2.53.20

Behind with revies and checkins. Sorry.

^reviews

and I am now posting latest working 2.53 source for my builds here archive.seamonkey-project.org/night…atest-253-nightly-source/index.html

thanks

Release Train

I actually have been doing ths for a shile but it seemed to be only intermitently working. I think now an be relied on.

no eta for 2.53.20b1. Should we do it soon because of the build changes or wait?

115 has been extended to march next year so artifact deps should remain available from mozilla and the pressure in replacing it being a time constraint is lessened.. but annoyance still pushes that research and such forward ;)

I am not sure the build changes are that important to anyone other than me, but not my call

only diff is i can now build central and 2.53 with the same toolset

might ease things with linux distros, but then I suppose these can be backported if needed?

maybe working out an official way to patch that linux builders can use but the mainline generic builds are done per normal

I have msys2 patches but they are broken. Early needed patch from 63 broke things.

i think we should be aiming to get 2.53.20 before Xmas

maybe even before end of November

oh so we have those who say Xmas should really be in spring so gives us more time

Hopefully, Microsoft isn't planning yet another mass-change in e-mail auth for 2024...

Yeah depends what comes up wrt security we could do a point release. Might wait with decommisioning CentOS 7 builder till then.

If I see anything regarding email and such I will report it to the channel

Extensions Tracking

bau

2.Next, Feature List, Planning & Roundtable

bau. Stymied with wip js becuase did need to take care of some other things. Hope to continue soon.

I am focusing on php stuff and infra more than my stuff in the queue patch wise except build stuff

I have similar issue with Tobin's pending patches trying to get those integrated into the central gitlab patch queue

to many personal issues taking up my time

well whatever the ones you are having trouble with I am pretty sure is a result of the environment I did it in and I should redo those parts from scratch

yeah, I'm working on various non-backporting stuff too

not many personal issues on my side, but I still end up not having anything for this § again, sadly

also got distracted fixing up my system to do Oauth2

hello world, late but whole~

I'm sorry IanN

nsITobin: no really just looking at the bugw you filed sommenting on them as I see fit and changing the patches to not bi TOP-CONTRIB- but <bug number>-

i have really not had time

hi tomman

AOB

oh is that a stones song? Out of TIme!

(next meeting will be last one in WEST, but it seems we will then dodge the period where US and the EU are out of sync in DST?)

rsx11m is out advisor...

*our

can always put the datetime in the topic

European Summer Time should end October 27th, so we are safe for the next meeting

need to leave early today. See you soon.

here? probably not enough room, didn't we have troubles with truncating before?

North America is November 3rd

well maybe I should use that site i have that I cna change on a whim and put that kind of info

My felling is need to get U in sync with EU. on both ends US is stupid is way to dark in spring for kids out on the street waitng for school bus and in fall some idiots (who i gusss never had kids) thout that elying it a week would mean kids would be out for Holloween wile it was earlier so not s dark. Real kids will refuse to go out Tiek or Tretink till sunset no matter what the closk says.

we have no summer time at Venezuela~

so the schedules are always moving here

EU was supposed to get rid of DST, but that move seems to be stalled somehow

timeanddate.com says EU switches last Sunday October, so it's still switching back

I'm again not sure I see a great need to move daylight a bit earlier, and I think that really only works for some latitudes, if it has a positive effect.

part of the issue is way back in the day when the us was smaller did not really need timezones so resulted in the eastern timeaone being way biffer than it should be and hard to change that.

anyway, thanks for your time today. Next meeting in 3 weeks, same bat time and same bat channel

on the eastern side should be in the atlantic time aonce and paty of the wertern part shuld be on us central time

and same bat UTC offset :-)

yepp!

yeah :)

I am from Indiana, I naturally hate DST

ugh i have to write so much PHP stuff

bye then, have fun!

websites, add-on site, pastebin, bugzilla-lite

I'm not sure I see the interest in moving the schedule if it gets dark just at most a couple hours after, say, an early lunch. I'm sure it's even more pointless in, say, Lappi. Issues fixed by lack of light having light will just move to the afternoon-evening, won't they?

nsITobin: is that not the stae where hafl was on eastern and half on central time?

i think the electric light basically eliminates most applications of DST

WG9s: at one point but not during my life

could be worse

-0430 VET

NEVAR FORGET

So I am older than you i guess. was an issue only in trying to figure out what time the Indianapolis 500 was

it's like driving in two lanes at the same time at a busy highway

tomman: oh now wounds like that Seinfled episode where Kramer adopted part of a highway and made lanes 2X as wide

nsITobin: Re: yesterday's IBM chat, I wonder if IBM didn't actually learn the lesson and separate what they had a very high reputation for from the rest. By selling the laptop business and keeping the rest :-P

possible

do IBM still sell things in 2024?!

Oh and for those who don;t get the same bat time same bat channel thing. Batman was the first tv show in the US that had 1 our storyies in tow parts with a cliff hanger was on conseutive nights adn after the first night was tune in tomorrow same bat time same bat channel.

WG9s: I doubt many are unaware of 60s batman

and if so their parents should be ashamed

but not that has shows one story on two nights

i bet some are two young to know

i don;t even remeber if iy was tuesday and Wednesdeay nitght or Wednesday and THursday

but and hour ahow cut in half and showed on consecutive nights to have a mid show aliff hanger

cliff hanger

I guess it wans't Fridays for sure, that seems to be taboo in US TV programming?

and then peole thought that was great so started the thing on last episode being a cliff hanger sontinued on next season (which I think was a ply to keep the network to not cancelling the show for the following season)

won't work great unless your show is named Futurama :-P

TRUE

so what's its current status? Un-un-un-un-un-un-un-canceled?

sounds like my projects

why do we live in a technicratic dystopian still somehow painted as pre-millinium life?

world*

because a lot of people dislike Windows Me

Oh ans now Just un=muted the TV and Dawn dish washing detergent seems to be using that Bonny Tyler song, Is there a total eclipse imminent???

it's 2024, we know how to tame the beast, I'm finding WinMe more and more charming every day~

reading a list, might be just un-un-un-un-un-canceled?

...unlike 25 years ago when I had... other thoughts

well when I read IBM talking about a holistic approches to AI sustainablity I think.. boy am I gonna earn that anti-citizen designation sooner than I thought

so... when SeaMonkey e.V. is pivoting from "browser suites" to "AI and advertising"?

I want my shares! :D

I'm not sure I want an autocomplete in M&N...

Github already revoked my implicit citizenship in their social network (and git hosting) service

you watch linkedin and github will cross-connect or merge

and THEN merged into the store

where microsoft will sell your apps and your skills

and keep 30%

plus ad revenue

The role of GitHub can be even worse, it seems some other projects rely on GitHub for identification, yet GitHub will terminate/suspend accounts if they're used only for identification...

njsg: this is what most people .. with a say.. want to happen cause people like me will argue with them

that's LITERALLY all it is

github is a wall to keep undesirables out

like every other courtyard that has been built since the as a Service happened

and for those who don'r get that last time a total eclipse of the sun in the Northeast US somehow this song was resurrected and became a hit yet again.

njsg: look at the hyper-amplified stories one after another about supply chain of software and security vulnerbility after vulnerbility shit that would have gone barely mentioned in years prior and just patched is being sensationalized more and more and its all moving towards not allowing any non-sanctioned ANYTHING anywhere

I love this song and am a great Bonny Tyler fan but this is not about solar eclipses.

and small teams cannot be trusted or expected to produce anything of worth because they aren't backed by a huge corperate entity.. its moved from indiviual developers to now small dev groups.. i have been reading

and it was ALWAYS pervasive in lunar orbit now seems everywhere

web of trust is another bunch of nonsense

that threatens to kill my ability to even run a website

I alreayd need bullshit worthless fake certs from let's encrypt for every non-good web client to even reconize as legit

but that's the CA system, not web-of-trust. I'd argue a major flaw in the CA system is that AFAIK it doesn't allow a web-of-trust effect.

well i may be citing the wrong scheme a few did show up around the same time

but its all

that

I'd agrue that https is only needed in cases where money ....... I figured it out

I'll be back later on.

no just don;t you want people watching what you are doing