Is there a reason why TLS 1.0 and 1.1 are still available by default?

I'm pretty sure the other browsers dropping them have forced them off the Web already.

1.3 should be default? least it was in 2.53.10?

in my 2.53.15 i have 1.2 an 1.3 default

It's the default, but 1.0 and 1.1 were both yanked in other browsers.

If nothing on the Web needs them and all they may be useful for is an attack on the user, they should be removed.

there.. NOW I am ready for the future personal.mattatobin.com/image/captu…cff-0fa5-409b-834e-8bbe14c0d61d.jpg

Okay, if I actually use Fedora's SeaMonkey 2.53.14 it messes itself up and starts crashing randomly after a while.

Like, from that point out until you delete the profile.

Okay, NoScript is the culprit?

"Updates to the following DOM HTML element interfaces: Embed, Object, Anchor, Area, Button, Frame, Canvas, IFrame, Link, Image, MenuItem, TextArea, Source, Select, Option, Script and Html. Please test add-ons."

Hmm. My guess is something happened regarding the extension platform and NoScript to where it's now crashing the browser.

IsambardPrince: Buc's 2.53.14 crashes on browser open on my mailnew profile, but not on my CZ profile.

IsambardPrince: only my mailnews profile has NoScript

IsambardPrince: I told buc about this

Well, at least I know what's doing it now.

does it only happen using buc's 2.53.14 or does it also happen on the official download?

a-865 and IsambardPrince ^^

IanN_Away does it only happen using buc's 2.53.14 or does it also happen on the official download?

I'm using the build from Fedora 37.

With NoScript installed, you'll be browsing along and then all of a sudden, any page you type in will start to load, then the browser hangs and the OS asks if you want to kill it.

brb

something occured to me

frg .. i am trying to compile nss newer than 3.64 with true vs2015 .. you guys are using the 2017 version of the 2015 compiler

newer vs might simply work

i mean that is partly why you went to the trouble to do 2015 on 2017 right?

IanN: didn't try official yet, but suspect it's NoScript caused. I use buc's because iso 8601 is missing from official.

there I reordered the xref sources

seamonkey is second, mozilla is third, moonzilla is dead fuckin last

frg when you show up.. do you want an xref copy of base-56?

need to add waterfox classic as well

this is EXACTLY the kind of service that should be on thereisonlyxul.org

ratehr than binaryoutcast.com

since I am offering it to others

cause even if my past wasn't a factor I am sure it is kinda annoying to link to a compeditor's product domain for a shared resource

maybe what we need is a XUL Working Group

fuckin take over the god damned specification for XUL and XBL

I'll offer SeaMonkey and even Moonchild membership

a-865: any idea if iso 8601 is covered in a bug?

need to add esr102 and regen final esr91

too

hi IanN

how are you?

I will likely relocate most of the cross-reference sources to a different xr instance under that domain I am paying for and am not using

put it on bismuth

CaptainTobin: just doing some sysadmin work on one of my boxes

fun fun

i should be sleeping

CaptainTobin: sleep can be good

CaptainTobin: though for some reason I was dreaming of cakes

IanN: I am just super motivated again

to do the stuff

AND THINGS!

wait until I strip the old Mozilla Page Styling out of the cgi output and wrap something nicer around it

switch the directory view over to PHP generated

while it is true the ident linkification DOES work in that view it is almost never used so fuck it.. get rid of that part of the perl man

i could take over freetext search results too cause that just calls out to glimpse

but again i'd loose linkification but that is ok.. ident linkfication is MOST USEFUL within source files them selves anyway

eventually i could likely get it down to just feeding the source file into the linkifier and taking the output

beyond that would require full understanding of the way genxref works

and port it to something else.. likely python

and crap out sqlite instead of dbm

that identifer search and the relation between cpp js idl etc is what makes it even today the superior tool for this type of codebase

heh.. I kinda wanna put the win2k3 sp1 sourcecode through it

i might shit a brick

there we go!

I need to add basic identifer rust support

cause in some isnatnces names were reused

hi WG9s and WG9s_

hi I

hi IanN

NoScript 5.1.9 definitely works with the official SeaMonkey 2.53.14. If it breaks because of a backport I will personally rip out this backport :) Would be great if someone could fork classic because it has some minor issues. I took out the embedded webext in my local copy because it would block shutdown for 8s. Only needed for backing up the prefs for the future webext version which...

...probably will be crippled by mv3 too.

CaptianTobin Support for VS2017 is in 2.49.5. I extended it to VS2019 there but only for the VS2017 toolset inside to retain XP compatibility. Because of spotty later standars support and some bugs left I can't recommand VS2015 any long and took it out of 2.53 sources. With 2.53 I am on VS2019 and clang. VS2022 works but is a bit too bleeding edge for my taste. VS2017 support is still in too.

Not sure if this patch made it into the official 2.49.5 sources: paste.debian.net/1255694

An mxr or dxr with annotation support would be nice but I can live with the current xr. Use if frequently to look things up. tortoisehg coves my annotation needs mostly and 50% of the time I need to check later versions for backports or future changes. esr52 with annotations would be nice for regression tracking.

a-865: Could you describe your "mailnews" profile to give me a chance to reproduce the crash locally?

frg: I tried to apply queueMicrotask patches, maybe this is the reason. :/

If a-865 has a reliably reproducible test, I would really like to reproduce it myself - then I can understand and fix this issue (of course, if it is queueMicrotask related).

But first, it could be fine to know whether the official build is affected as well. A big request to those who reported to check this on their own, that is, in your environment where the error manifests itself.

buc: IanN frg_Away I copied my sm-mailnews profile, hid the original, used the copy to open official 2.53.14, and it did not crash on (default browser window only) open as Buc's 2.53.14 does

buc: describe how, other than as in my email?

buc: 129 files in chrome directory

buc: should I email you its prefs.js?

a-865: I'd like to have some steps what I should do on a clean profile to try to reproduce this error.

a-865: Try to strip private etc. data from the testing profile, strip out some non-affecting extensions etc., and then I could try to reproduce it myself. I beileve it will be faster, rather than trying to decsribe things

buc: disabling NoScript while running official, then starting yours doesn't crash

buc: enabling NoScript while running yours causes coredump

a-865: Could you please first try my new build buc.fedorapeople.org/seamonkey/download/seamonkey-2.53.14-2.tar.xz ?

buc: segfaulted :(

a-865: Well, then it is not queueMicrotask patches related. :(

Very thanks to a-865 for providing test profile with reliable reproduced bug!

The issue goes away when I remove "queueMicrotask" patches, which seem uncomplete for now.

For me, it crashes even with one opened tab with avsforum.com/forums/lcd-flat-panel-displays.166 (or another avsforum.com page)

frg: If you have a build with qM patches applied, could you confirm that the link above reliable crashes? (NoScripts seems needed too)

buc told you best not to apply them when I asked you to test. I don't have a build but with these they crashed for me on another website every time so it seems like the same error. It looks like some recursive call hits an assertion. I debugged it a bit then but was unable to find a solution for now. Trying to apply some more stuff for Shadow Dom first because this is where they didn't apply...

...clean.

frg: Accroding to backtraces etc. looks like bug 1443746

but first, like bug 1443429 (the same final backtrace fragment)

Interesting, that applying bug 1443746 only fixes the crash (some trivial, but generic event fix)

buc cool. I thought it was something like this but didn't find the culprit. Looks like a kludge but not the first one.

IsambardPrince: buc made my crash go away

I am in the middle of a big rebase so won't make todays 2.53.15b1 pre

Also, while I'm here anyway....

Palefills stopped working at all

I was looking forward to my Google Drive working again.

IsambardPrince: martok/palefill #48

just got bitten by that yesterday, but found a easy fix

Will there be a bump soon or is there something I can do locally to work around this?

brb

IsambardPrince: it's a one-line edit, but of course you need to download the XPI, unzip, mod, repack, and reinstall

the issue hasn't been acknowledged yet, but then I filled it on a Saturday night :D

tomman: thanks for filing the issue and working out what broke and needs fixing

what does having polyfill installed and working accomplish?