i need to rebase them against the latest patches but here are my cc patches code.binaryoutcast.com/thereisonlyx…-patches/src/TRUNK/mzla/comm/series

if you care tomman

welp guess I am doing bugzilla.mozilla.org/1912236 on central tomorrow

too late tonight

Services.tobin.sleep();

couldn't sleep so I did this instead.. simplified xulorg's template .. still have to adjust the color of the stolen navigation sidebar thereisonlyxul.org

Good I am happy with thereisonlyxul.org's modifications.. makes it easier to click where I wanna go all in one place

good morning #SeaMonkey

nsITobin greetings

ah hello frg_Away

well frg i have VMs to build wip and central and I have a repo to store mah patches in and i have patches to finalize and i modified xulorg's design and front page so i am pretty happy atm

cause Real Fun(tm) shall commence shortly, I'm sure.

nsITobin given that IanN has not objecxted to bug 1913787 might wantr to check the Firefox menu. Seems it is using history. I would still keep the sync item in it even if it is unlikely to be activated again in this life.

well issue is it was mangled into the go menu

would also keep the first four go items

this would be resolved with a syncui and backend replacement tho

otherwise I am gonna have to split it out to its own function.. and then comment it out and deal with it later

personally I'd rather rewrite the previous patch with ifdefs so when/if it gets fixed/replaced/whatever its easy to find

maybe into tools? Not a big deal anyway. Or do a seperate part 1 removing it so we cann easier restore it later nor never :)

likely navigatorOverlay.js

where the menu is defined

in the xul file

I would THINK

splitting it would be better for ANY contengency rewriting the older patch or just backing both out.. ye

I will do that

nsITobin: wondering what it was you thought you needed to do for bug 1912236. Wanted to make sure I already allied this to suite in the 2.53 gitlab patch queue wg9s.com/comm-central/patches/seamo…999-port1908725-suite-bustage.patch

Ok so you know I have this basic browser buoy window and I use a browser element and thunderbird's MailE10SUtils to manage it so i can have basic navigation .. due to this bug the behavior is it spawns a new process and new main window instead of doing stuff in tthe browser element also doesn't rememeber what it is supposed to do

that is how I came across it

once I realized it was spawning new processes i knew e10s so i search thunderbird

reading that bug was uncanny tho WG9s .. they do not know the history of their own project

even those that SHOULD

okay so I already have this patch in my working queue WG9s

so not sure about palemoon but for wuite I needed to add the moz.build change for jsshell wanted to make sure you was that becuase if you do the TB fix in the bug you referenced will reuslt infailured bulding js shell under windows

trying to make sure you don;t ave to reporduce stuff ws already a solution for

what does Pale Moon have anything to do with it?

i did not know why you thought needed to work on it becuase i had alredy landed the fix on central gitlab patch queue several days erlier

i didn't realize you had ported it

i guess it isn't enough

so thought perhaps working on a different project build off of comm-central

cause control over browser elements is busted

WG9s: at this inital point.. its all the same work

well se that is wone of the issues where i have only been tasked to make sure the suite still builds and since the UI is terminally broke is hare to test anything

getting SM to be able to test stuff requires a working browser element that is controllable as much as an indpendant webrunner does

OK was confused i figured you either had not seen my patch or was working on a different project (like PaleMooon)

I didn't see it

seems you re past wat i have. do you actually have patces that make the UI work to any extent? if so I would love to add them

I only refreshe dthe patches last night and built

my state before that was before this landed

The current status (before this bug) was SeaMonkey could build.. and launch into an alternatively created main window with devtools and a working --chrome command line handler added to devtool's that is where I locked it off

so all i did was prot waht seemed to still aplly form the broswer ptches in te initial bug then add stuff not already included and changed in the TB patch and then add the js sell stuff to add the propsys library

even without my temp ui just the devtools with chrome cli flag is immensely helpful for testing

but how do i get to be able to run devtools if the ui does not come up?

seems you know some magic that i am not aware of

well the reason the UI doesn't come up is because we have no functional (and later no) glue that tells the thing what to do with its self

or have patches applied that i am not aware of

WG9s: trenchxulfare

i don't recommend it

lol

WG9s: when we last spoke about this i was just creating and figuring it out

but central is secondary to 253 so i don't want to get lost in the wilderness they planeted after bulldozing all the useful stuff

if you have patches that help and do not break the build send them to me and i will add locally to my build and if they loook good I will try to get frg_Away to approve them

I am actively getting them organized and so they can be added on top of current patch queues that is what I was working on past few days

I didn't wanna make TOO much of a fuss until i got it good and not changing it every hour lol

well just saying of they make thinks better on my builds frg_Away would probably love them to be added

not sure what platforms you test your code on.

WG9s: these need rebased and cleaned up and headers fixed but you can see what i gots code.binaryoutcast.com/thereisonlyx…/brineshrimp-patches/src/TRUNK/mzla

central

just checked out last night

also put this up yesterday files.thereisonlyxul.org/version-control/hg

more valuable than later version tarballs

Os so clean them up and fix headers and wend me a zip with all the patches and a series equiv type file to show intended order to apply the patches and I will start incldeding them in my central nightly builds, shich I think would assist in getting the ptches to be accepted

WG9s: if you had waited 24 hours the issue of not knowing what I have or have been doing would have been resolved ;)

and depding on what i think, i resrver the right to add them to the gitlab central quwue without frg_Away approval. Afterall, I do that with my patches.

i do my queues with a series file and subdirs so that it can simply be "overlayed" on top of the sm queues and series file contents pasted in

so its far easier for others to consume

figured it out 2 days ago lol

i think its a good idea

so you have a site somplace with new patches and a series file that I could apply on top of my crap?

I usually check in everything which breaks building and leave the others in the queue so that future checkins are easier. Usually it takes much less time to rebase the queue instead of rebasing patches.

I don't have em up on a naked http dir no..

I might prefer to wait till ou think the patches are worthy of landing on the gitlab central zueue

these should be ok but they need headers revised

i have 2 states of patches similar. those I apply locally as part of my build script and those I land on gitlab so as to be expected to be added to everyones build

null-dll-blocklist is a top patch not for central

unless they accept my platform patch which they won't

WG9s: I am just copying and pasting back and forth moving my gaurd fake-patches in place.. extrapolated from what frg seems to do

i do need a pass to fix the headers tho that needs scripted like you have mentioned WG9s

I am seriously getting into this development style i kinda like it

what is wrong with headers? all i care about in headers is that they have a valid hg header line. A user line that contains a valid contributor name with a valid email address and a valid Date line

not all of em do cause not all of em have had an export pass ;)

something I am working on

i don;t like to put things on gitlab that will fail when the convert the Mercuril prop to git thing they do at release time. is fussy about user like without an email and either no or unparsable date format

swap yours for this code.binaryoutcast.com/thereisonlyx…WIP-1783623-port1524687-suite.patch follow up with

and do these anywhere

that SHOULD get you access to spawn any window from commandline as well as an attached browser console by normal means

otheres seem to care about NOde ID and Parent ehaders but as they do not break the port HG stuff to git thing I don;t really care.

code.binaryoutcast.com/thereisonlyx…ib/null-dll-blocklist-bustage.patch and code.binaryoutcast.com/thereisonlyx…ISHED-unrestricted-comm-build.patch

so it can run

and not crash

the rest is a mess

or busted now with that bug

WG9s: I am not sure it actually matters in mq

the node id and parent headers

but i bet bugzilla likes em

exactly frg_Away seems to like to change them my feeling is no one cares

but I will do them as you guys want because the idea of contributing is not making extra work lol

because they are likely to depend on other work in local queue so incorrect anyway

for nayonone other than you (or me on mine)

I try very hard not to make those mistakes.. but no promises ;)

WG9s: this little patch repo

also has another secret

it's unfinished but helps

heh need to update the bundle baseurl

ids are recalculated during qrefresh. Doesn't matter. I only check ocassionally when I backport with the original bug that I have the reght revision. Afterwards new or old doesn't matter.

my intenion with this script is to basically automate a simple way the task of getting wip (and 115) or central, running bootstrap in 115 or central (in central's case change the origin url to mozilla and pull) and then patch.. and after ALL THAT.. Then apply my patches ;)

I also want to attempt to figure out a way to rewrite the visual studio finding logic to pick up the cached 2019 version that 115 downloads if it is there otherwise exclude it and download and run vswhere with a predefined configuration

so that the entire process is minimal hassle even if it will take quite some time

among other patches both 253 and central and other BinOC things too like an operating system lol

I must do more.

WG9s: there is a simple cheat to ns*App.cpp

copypasta

and change Thunderbird to SeaMonkey

nsITobin: but would be great f we were not at cross purposes. WOuld need to treat the central hg repo plus the gitlab seamonkey-central-patchs repo as the base so that we are wriing patches based on te same base.

and if you think that might work i will try to get you patches landed on gitlab based only on my approval

so like if i think this breaks nothing and maes tings better i sill wuickly approve and add

we seems to both have the same goal so no reason to not be coorperating

WG9s: does it make sense to still consider central as future seamonkey 253?

or maybe I am misunderstanding

frg has gen up on central but maybe not adverse to geting it ot m-c68 esr

so still possible issue is he stillwants to keep wupporting legacy add-ons

WG9s: how can one build esr68 tho i couldn't get it to do it because I'd need to manually solve deps because taskcluster no longer cares or provides for older codebases

and i think many don;t get that that is a huge security issue. is an API that add-ons are supposed ot use but nothing prevents a legacy add-on to bypass the apis and just access global variables to make completely arbitrary unsupported changes to the code.

WG9s: there is no reason to believe one cannot have XUL Extensions at ESR68 just need some reverts it just may not be the full spectrum of restartful toolkit extensions.. they would be bootstrapped unless changes were made otherwise.. THIS is an area I know very well the jump platform upgrade and revert thing

esr 68 makes no sense. missing tons of later js stuff and legacy add-on support is so gutted it would basically all add-ons to be changed.

so the mozilla firefox pi\eople tried to do this by only allowing add-ons on AMO that did not do this, but then nothing preventing add-ons form other sources

other than a warning

so have to overrride that to tet ad blocking to use and once you do then anyone else with a malicious add-on is also permitted

WG9s: I disagree with you on the security issue of having powerful application modifing extensions and here's why.. Properly informed and educated users with approerate education and informational resources are more than capable of deciding if they want or should run packaged app hacks and the consiquences if it goes wrong.. This is what killed the technology.. lack of proper user education .. and a willingnes to have a powerful open extension system

open by default in the 2010s when most don't need much as "proven" by webextensions.. it should be there but opt in.. a power user feature for power user modifications.. but this again is only a problem because of Mozilla mistakes and sabotaging by dumb decisions the fostering and curation of this technology and ecosystem

also i want them too LOL but i also want binary xpcom extensions on top ;)

a whitelist of add-ons that are permitted without having to get them working by allowing all add-ons this priv woud have been way more appropriate.

they tried this but not sure what happened and decided to just not allow legacy add-ons at al

I'd be fine with babymode computing as the default if I had a quick and effective way to make it go away so I cna get back to what I was doing that is my problem with modern security

Modern Security is drenched in the assumption that I as the user do not know better than the machine

user.. OPERATOR

System Operator

I think your (and my) issues are the same people wtih no idea bout security trying to dictate what is best for security reasons.

if I question it.. maybe I can make the decision my self?

Privacy, Security, and general risk assessment can never be one fits all

nsOreason i said go via me for now is I am authozed to change gitlab central aptch queue and am not sure you are..

if I want to run Windows XP as NT AUTHORITY\SYSTEM on an unfirewalled open internet connection.. I am prepared to mitigate or face the consiquences of that action. Even if I admit its kinda stupid.

and make a very huge understatement ;)

but I still think betr if we work together

I don't want anymore access to anything than you and the others feel is approrate and required to faciliate activities.

nsITobin: not up to me

i don't think I need anything currently and i might mess it up as I am still leanring and forming my own workflows based upon em

except to document what and how and whatever i am doing better

maybe start a blog

LOL

so its all.. well what is tobin doing *looks* ah well this and this i want this is crap and this needs to be rewritten AND WHY are there still BinOC patches lol

just sayin if you ork from same base as me easier fo rme to advocate for you patch if you anet me to build using teh current hg and ignote the gitlab patch queue and use your's instead (which is what I think you said) not so easy fro me to help you here

I can stay in sync

I just didn't do that initally cause its INSANE on 253 atm

I knew you wul understand

try forming it as an order

;)

if you work on the same base as me i can easiy test your patches on allplatroms

seriously yeah I didn't mean to split off but that last bustage and this one has kinda thown a wrench in my central queue

cause I was working on prelem es conversion by hand

but changed gears becuase that is gonna be a time sink and i need to get stuff where you guys can easily access it instead of throwing links constantly to adhoc files

nsITobin: not making it an order just saying fi not forking is wasier to get it landed frg_Away wirites patches for central to include m bustage fixes. he is a way behind of this becuase of his other issues, but is now back to all my pending bustage patches will end up on comm and mozilla central so would be better if you dealt with thm now

but also if one of tem looks completely bogus let me knwo and i might remove it if you have a better replacement

well I got the VM for central up

also shuld havw been py8ing for attention. my requiremnt meant did not require a rull UI but required at least a small window with an X choice to kill the app

seems that had not worked in a log time. I should ahve been apytin more attention.

WG9s: I been chaotic with false starts for over 2 years now.. i don't blame you for not paying much attention until stuff materalially starts happening

the ONE damned thing that is annoying about working with patches and it isn't mq's fault is popping off anything configure or a configdep means i have to wait an hour

kicked off with a whole code transplant from Thunderbird's nsMailApp.cpp to SeaMonkey's nsSuiteApp.cpp (with a few strings replaced)

any bets if it will work, work as intended, or fail

i have no ide

idea IanN was last to look at this code from the suite perspective

frg_Away: prefers changes to be bug by bug changes and not a wholesale replace, just sayin

nsITobin: besided ust replacing one .cpp file with a newer version without looking at the changes this actually had done on central in defined version changed other files that your just replace this does not do so this is more likely to cause more than it fixes. Just my opinion.

WG9s: if i reverse the clang style changes it will be effectively the bug

oh ok then fine

xre bootstrap made ns*App.cpp almsot identical across the board

but still check with frg_Away sometimes he is fine with jut doing the style changes other times he prefers dong the changes with out the style changes

indeed

the downside with --jsconsole and --chrome vs entry points from Buoy is you only have the life of that one window and since everything is busted you can't do much else.. that is why the minimal window I created as a substute entry point for SuiteGlue effectively and as an isolated working (theoretically) thing can let you fix what you want if you want or at least pull something up in the browser element

WG9s: its a crime you won't find ANY of this outside deep in the shitpile of the Pale Moon forum

or this channel's log

the real js console was more useful for this as an ancor

but that is long gone

WG9s: did you add the followups? maybe that's the issue for the difference

yep I think that is it

why didn't they squash this before landing proper

it is my estmimation the last real codebase to have something akin to a XUL platform would be esr102 with heavy reverts and retroengineering.. beyond that there just isn't enough mozilla left in Mozilla beyond the rendering engine and accessing it is a pain in the ass these days but I will find a way for one project or another

I wonder how bad thunderbird reacted to this.. i bet there were half a dozen main windows open all not right

WG9s: confirmed

needed followup patches

nsITobin: so if you do those either one bug at at ime (i am not a sticler if more that one patch in a bug (ust try8ing to satisfy frg_Away requirements.

IanNs too :)

it was easy to miss the last follow up landed a day later and without having anything to actually display and see how could you have known anything..

When I did the library replacement it was almost impossible to do this bug based but that was/is one of the few exceptions.

frg_Away: do you care if nsBrowserApp is clangformatted?

err

suite app

I usually do manual formatting. IanN is sometimes allergic when it comes to my reformatting in patches so I try to not overdo it.

Anyway one day we need to do it for all files.

see while clangformatting moves the braces up like I like it also makes all the function arguments a messy list instead of listed out unless just a few

well from experience thunderbird's nsMailApp as-is with just a few string changes is a very effective and almost universal match for anything the suite is doing

especially now that most of the code was shipped off to xre bootstrap

in UXP everything was ultimently based on Basilisk's refernece when such a cheap hack would do ns*App, shellservice, other things.. aboutRedirector tho.. Thunderbird

nsITobin I usually move the braces up and leave the function alone. Current code looks all messy in mozilla-central

frg_Away: which formatting do you prefer? netscape, late mozilla, or clang

not that I am the greatest at ahearing to it I prefer to code in what I deem a updated netscape style.. which is different than late mozilla tho some earlier styles are still present in this PHP code

i need to fix a lot of alignment

.. mass renames

how mozilla of me

nsobin probably like a mixup too. Around mozilla 68 was ok so maybe clang 7 8 . The late formatting is bad

i'd seriously consider doing some pathfinding with esr68 but I can't build the bloody thing closest thing to it i can build is..

wait for it

SeaMonkey ;)

of course you know frg_Away .. having a patchqueue would have been invaluable to how I managed to achieve vNext 1 and 2 on Pale Moon

it was lots of branching long recommit times and then reapplying everything and doing a little more and over and over again up every mozilla version.. just aiming for and you guessed it.. simple buildabiliy first and foremost

esr68 would probably build with an old rust version and clang.

how do I downgrade clang

i downgraded rust but it failed

it's bugging me cause the xp people can do it

I shall not have the capabilities of Windows XP users exceed mine! lol

nsITobin you now and then grab artifacts from taskcluster. I thinki have clang 9 and 14 for Windows.

are they still up there?

That s how we cross compile for macOS

but basically same for Windows.

i have some old slng versions tar files tha you can install in ~/.mozbuild

and then specify them in your .mozconfig

WG9s: we should put it up on files dot xulorg

for sure

nsITobin not sure. Doubt it.

if have clang 11 and 13-17 in additon to the version 18 that mach bootstrap installs

only esr 115 and up on treeherder treeherder.mozilla.org/jobs?repo=mo…e49879c02d0c0af264f2fc3aafc07823782

so you can't find them via this route.

then in my bmozconbfig at top of the file I export CLANG=$home/mozbuild/clang and then do everytong else like CC CXX etc uisng CLANG

right i was searching for taskcluster

re-bookmarked

if you can;t find what you need elsehere i can email you the tar file

WG9s: I sent you creds to upload stuff to gallium if needed

please don't fill the disk the logbot doesn't like that ;)

gitlab wip updated. If anyone misses nsIDOMSVGLength aka bug 1435138 let me know.

will move it over to 2.53 next too.

frg_Away: trying to think what outside the machinery of website rendering scope would use it for

nsITobin and then there are also sw´vg web interfaces too.

well i just dunno what use svg manipulation has in an extension

unless you wanted to make an svg creator

nsITobin anything which every used it is probably so dead no one remembers it. And new stuff can use the regular interfaces so no loss I assume.

well of the popular sub-thousand extensions in my head .. but in a mist of sorts.. i can't recall any of em being anything svg related

my small brain neither.

firefox extension mist is not a typical brain condition i trust