hi frg_Away

Hi NewTobinParadigm

did you give any more thought to what would be useful to you on xr.binoc?

know what I like is that searchfox seems to explicitly exclude codebases Pale Moon has been on

24 38 and 52

but that's ok.. and even though I am not with /them/ anymore my cross-reference will continue to have a wide selection.. because well.. i can lol

Not much. Was deep into releasing the next beta with IanN and doing backports. I have "arranged" myself with the current stuff mozilla provides but as stated would like to have the latest 2.53 beta or relase branch from gitlab. I can live with dxr. I occasionally miss the older branches but annotate or local history with tortoise hg usually is ok.

i have the release branch but you are branching with different branches every time so is there some sort of feed so i can switch it out and reindex when needed

if you HAVE a branch i can just put in and have the script do it automatically like a beta or master or trunk branch then that can be automated

comm-sm2.53 is currently set to 2_53_11_1_final with the reletive mozilla srcdir included

NewToBInParadigm IanN is usually creating the branches. I usually populate them when we are all ok with the patch queue but a bit later. It is a bit of a mess and don't think can be automated. 2.53.12b1 is done and will be set when we release it as default

wonder if there is a way to query which branch is default from gitlab

git symbolic-ref refs/remotes/origin/HEAD | sed 's@^refs/remotes/origin/@@' seems to work but i dunno if that will stay valid if you change said default branch

I would however strongly suggest you go back to a trunk and release (tagged) or trunk and relbranch style at some point

I'll file it away until your next release and test if it does that and if it shows correctly i will add custom logic then I can.. SET IT and FORGET IT .. like a Ronco Showtime Rotisserie!

as for beta branches.. i dunno since there is nothing to query i don't think i can do much without staying on top of it.. unless you simply wanna provide a file somewhere with the branch names I can just curl down

I dislike this as much as everyone but using mercurial queues makes backporting really really easy. We can shuffle stuff around and fix it if errors are found later. Saves time and this is something we don't have enough. The l10n repo is basically traditional.

well if you prefer hg maybe look into some of these free options mercurial-scm.org/wiki/MercurialHosting

NOT SOURCEFORGE lol

If it helps I think putting the current branch name into a file in master is something easy done.

foss.heptapod.net/seamonkey is now basically the base.

Reminds my to sync l10n with gitlab :)

like a release-branch file with one-line on each would be very helpful

treat it like version.txt (no new line please)

IanN any preference for a file name in gitlab containing the latest branch version with current source?

what I need is a list of corrosion identifers so ident can search it

have to do that one day

"The Binary Outcast Cross-Reference Tool is a PHP executed hacked version of the Perl MXR tool which is a hacked version of 20 year old LXR."

yes.. I wrote in php a way to execute perl cgi

the reason is I am VERY SLOWLY moving functions to PHP

the only thing I won't be able to move without a bunch of work I don't have time for is identifier search and source file view (because of identifier linking).. dir list view, free text with glimpse or something more modern, and find file is basically trivial and i will get to it eventually

but i have a lot of OTHER infrastructure and projects to manage so.. that takes priority especially since everything I personally do I am doing basically alone now

BTW I am gonna revive and fork InstantBird for some reason

frg_Away: release.txt? source_repo.txt?

latest_branch.txt?

IanN latest_branch.txt sounds ok.

what about relbranch.txt and betabranch.txt?

cause you wanted beta as well right?

I can live with only beta /latest but yes can be split

which would be more useful?

beta or release?

or you just want what is effectively seamonkey-central just with this crazy branch setup right?

which is fine by me

and that means less logic to code up

I think having beta would over complicate things

well cross-refernce code search for current release is less useful for day to day dev than cross reference for current trunk or beta whatever your dev branch is

i will do both if you want that isn't a problem just need to know what would be useful to y'all

Usually few changes between beta and releae. I would just go with the latest branch be it beta or release. That is the most current in gitlab. so 2.53.12b1 2.53.12 2.53.13b1 2.53.13 and so on. A latest_branch.txt would be ok then.

what would you like it called? seamonkey-central sm253-latest?

central is still comm-central and would be confusing sm-253-latest

back in a minute

is seamonkey-latest253 ok?

kinda have a naming scheme going :P

seamonkey-latest2.53*

is seamonkey-latest2.53 ok?

kinda have a naming scheme going :P

didn't crash :)

I'll also add comm-central right now to the list

ugh comm-central doesn't have a proper git mirror does it

If you want to stay with the old names maybe comm-release-sm56-latest and mozilla-release-sm56-latest but overall pick a name and I am ok with it. Other than central don't think anything elase is confusing.

bit verbose

but no for your repo i was gonna have it index mozilla along with comm

tho comm-central and mozilla-central are seperate

see ;)

also that version is fake lol

cause it is a modified m-r56 in reality

Yes it is better

yeah since you changed the mozilla version you can't go back to what it really is .. that would cause issues with extensions adapted already and might even piss off aus

oh well

but there IS a more proper way to do it if you need to bump the rv version

in a later codebase

it has been a while since I used hg

is it hg pull or hg checkout?

Yes I wanted to set it to 56 back a few months ago but then web extensions did choke. We still build Firefox from it for tests and a-b testing. So 60 is it for now. Actually basically all of the 57 changesets which are not too nasty are in.

ok hg pull is effectively git fetch and hg checkout is effectively git pull

I thin hg and git clone are more or ess the same.

but statyed with 56.x for Thunderbird there.

Good fro a-b testing too

hg pull should be the command

ugh i simply can't remember

i think i am confusing my self remembering client.py

hg pull

 

Pull all new changesets into local

yeah

ok

Install Tortoise Hg. I use the command line but the gui is a lifesaver. Shows you the commands it executes too so you can pick them

heh well this is on a linux server so..

i haven't ran tortoise hg in .. a decade?

Available for Linux

gui integration is only available for gnome and I am on kde but the gui itself is fine there.

I don't run a GUI on my servers

Yes for automation it is not the tool. Just if you want to work with repos.

oh man

Pixiv went full Chrome

Google WebComponents® are now mandatory beyond the login page!

ridiculous

Japan was the last standing man on the pointless Chromeization of the web

...I guess that ended right now

to be fair Pixiv's UI was laggy as hell since they drank the SPA kool-aid, but they just went full-in with WebComponents® in the last few days

CLOSED WONTFIX USECHROME HEALTHYECOSYSTEM

frg_Away:

we will know in an hour if my updated script works

should

i still need to make it so it doesn't do everything every 6 hours but the comm-central and hg part is done

still think i am gonna have to make a major modification to get it to conditionally re-index so i will save that until i get the sm latest going good

likely will ALWAYS run for c-c and m-c

cause way too commit happy

hmmm, enabling dom.webcomponents.enabled is enough to "fix" Pixiv, but of course that can break more stuff elsewhere

tomman try the latest 2.53.13b1 pre. has the patch from buc for optional chaining in. Together with Nullish coalesce op unbreaks lots of sites. Depending on what pixv uses setting dom.webcomponents.customelements.enabled and dom.webcomponents.enabled might work somewhat now too. At least gitlab and github are almost displaying ok with it. Setting them permanently to true not recommended.

wonder at which extent Pixiv is using WebComponents®

frg_Away: awesome~

please excluse the inital commit message.. after what happened i was very angry and cross-reference was one of the few repos i didn't have a local clone of and lost but the code for it lives here.. code.binaryoutcast.com/infrastructure/xr

Maybe like gitlab initially for emoji support. Core stuff for sure...

patches welcome

frg_Away: just enabling dom.webcomponents.enabled lets Pixiv work (of course I'm aware that it's a global switch and enabling it all the times can lead to Bad Things™ happening)

In fact.. I am gonna force push a revision

I guess it's yet another case of "our 3rd-party framework just pulled that on our last point release"

fixed

will try .13b1 as soon as my DSL stops wetting the bed, has been working terribad since last Wednesday

frg_Away: I likely have more webcomponents than 56 does but yeah same deal.. it is literally an all or nothing evil spec

I am at about 59a1 with custom elements now. shadow dom is a bit behind. stylo would be working in theory if it wer not for the rust compiler crap. Old version does not work any longer and new version needs new bindings code. But unless 10 devs show up and we could remove this crap it is mandatory unfortunately.

what I would like is to have the power to force the entire Internet to stop using Google WebComponents®

I've yet to find a single developer to tell me what's so great about that crap, aside of "well, it's modern"

I'd say aura (and uxp) is somewhere between 62 and 65 with some parts older some parts newer.. we do NOT have dynamic es modules but do have basic es modules.. it can't progress further without ::slotted and another one i can't recall off hand

who killed server-side frameworks that we now must push page processing client-side?!

and because it was stylo there was no nsLayout version to work off of

tomman: well thanks to Moonchild loosing us the browser war .. it ain't likely

well that and Mozilla surrendering

I blame webdevs rather than browser devs

As usualy the theory of using components is fine. In practice only used to enrich google.

webdevs never asked for this

it's the same reason of why Mac sells

they didn't ask for hardly anything of the past decade

this is ALL google engineers

and sycophants

and Mozilla allowing it

and webdevs jumping into the fray because "oooh, shiny, must profit!"

real webdevs don't use this shit..

why would they

too bad "real webdevs" are like dinosaurs, they're largely extinct these days

nowadays everybody relies on multi-megabyte JS frameworks

Eich said js was an abomination when he wrote it

and if he had not been purged for disliking the gays getting married maybe Mozilla wouldn't have surrendered so quickly and Eich wouldn't be a chrome shill using deceptive practices to temporarily steal money from people under false pretenses

his purging really broke him

in the meanwhile, even cellphone software is nothing but more Chrome-in-a-can

magic slabs of glass have done more damage to us as a species as anything else

In any case, I'm still waiting for the first webdev that can give me a honest appreciation of why Google WebComponents® is so great

what you can do with that horrible spec that you can't do with vanilla JS (or even better: server-side!)

heh, and people DARED complaining about J2EE/JSF by being bloaty horrible enterprisey crapola, but IMO it's saner than anything JS

WebComponents at its core is a twisted version of 20 year old Mozilla concepts that were sabotaged by mozilla them selves 10 years ago

prove me wrong

didn't even MS tried that two decades ago with HTA?

(and failed miserably)

HTA was never XUL and XBL

and Mozilla thanks to their arrogant 2010s bullshit sabotaged their own mature technology only for Google to reimagine it in the worst possible way years later

if Mozilla could have solved remote xul's security issues.. we'd be golden and it WOULD be the defacto standard for webapps

modern webapps

the only reason it didn't happen was because it wasn't allowed to

xulrunner, remote xul, et all would be king over anything Microsoft would have come up with and Google would have had to support Mozilla technology

not electron and webcomponents

or xaml/directui

I do not want webapps everywhere, that's for sure

I get it, native development is hard (and another can of worms), but it exist for very good reasons

stop trying to turn the browser into a full-blown OS!

even cellphones have to endure endless layers of bloat because every app is another Chrome-in-a-can™ of unusable bloat

well with javascript being TURNED into something more complex and ridiculously opaque as any native binary especially with webpackage (webpack?) yeah

(not everyone can afford an iPhone X or Galaxy S)

boot2gecko was never gonna be successful

and kaios still isn't

xul fennec was screwed over

I own a KaiOS phone because I'm masochist, I guess :P

i have one as well

it's as terrible as you can imagine

and with 256MB RAM, well, ouch

it is more terrible than I imagined

it is like whoever works for KaiOS has never used a cellphone in their lifetime

my 2004 feature phone runs better

there is LITERALLY a 1.5 second delay between keypress and action

how in the hell you come up with a phone UI where rooting is required to change the SMS ringotne?!

it is pathetic

reminds me of old Verizon cellphones where everything was locked down behind crippled firmware

but at least those were fixable

anyway.. if you wanna see what a communicator PLATFORM codebase looks like keep an eye on the Aura Runtime Environment heh

I am spreading netscape design throughout the toolkit!

xpfe design will return!

next to me there are three cellphones on top of an old Pentium 4 laptop: a KaiOS turd, some cheapo Snapdragon-based Android 7 Alcatel

and a fresh out of the box Redmi Note 11

i miss xpinstall

...all of them are terrible compared to the P4 laptop!

i'd KILL for someone to reimpl full xpinstall

why did Mozilla kill that technology

a cross platform installer tool common to install, update, add-ons, components

it was absolutely stellar

forgot the "monetization" part

i am just eBegging

that IS my monetization

you need VC funding, the IPO, unicorns, and just to be safe, crypto/NFTs

;)

also, getting Musk tweet about your stuff always help

good luck making THAT happen

if Musk tweeted he used SeaMonkey.. you could TAKE OVER comm-central and fund a whole infrastructure and buy out the trademark from mozilla

it's the same guy that shot Dogecoin near the $1 point (and back), and that considers Touhou wolfgirls to be catgirls instead

so weirdest things have happened :D

I'm a Dalek.

but you knew that already

gitlab now has a latest_branch.txt in master:

Need to do some real life stuff now

frg_Away: I'll get this done today just not this second

2.53.11 isn't displaying the whole page, missing most of right column, and missing comments following the first: gitlab.freedesktop.org/drm/intel/-/issues/5711

what does the Toolkit Error Console or DevTools Browser Console say a-865

Timestamp: 15/04/2022, 15.50.55 GMT-4

Error: TypeError: e is undefined

Source File: acsbap.com/api/app/assets/js/acsb.js

Line: 1

is that the only error? also perhaps pastebin em?

this is constantly repeating

only error I get is invalid regex group but my codebase is older than sm-latest

i don't have seamonkey available atm

there are others, but the window scrolls so fast they disappear before I can copy...

closing the tab leaves the other gitlab tabs doing the same

try the browser console

won't be as nice imo but shouldn't flow away

as fast

I tried copying one field to a text file to paste. Paste fails, 237k bytes O_O

constant stream of errors from all over the place

doesn't GitLab requires JustOff's addon?

you need an OLDER version of the addon because lolGiggitylab and lolFreeDesktop

...except that nope, 1.2.16 won't help

...but updating to 1.2.18.2 works

(I had to open a fresh tab)

now I can read the whole ticket

SyntaxError: invalid regexp group commons-pages.admin.topics.edit-pages.admin.topics.new-pages.groups.milestones.edit-pages.groups.mil-2c3731d7.9d206744.chunk.js:1:285

got that from browser console after reopening closed gitlab tab

a-865: you need JustOff's github addon to work around that

make sure to install the latest 1.2.18.2 release

bugtrackers ARE supposed to be solid, simple, reliable software with as minimal dependencies as possible, so why do they need emoji regexps!?

WHY

> Does this page need fixes or improvements? Open an issue or contribute a merge request to help make GitLab more lovable. At GitLab, everyone can contribute!

the addon works

where "everyone" means "everyone using Chrome, of course"

tomman: anything to bloat their pages more :p

also, seriously, emojis ON the developer console!?

soon, we'll have emoji-driven development

i still intensely hate JustOff

btw

that will never change

attempting to open the invitation to new issue opens this instead: unagi.amazon.com/1/events/com.amazon.csm.csa.prod

that is why I turned his bootstrap extension into a jsm and am building it into my navigator

it just needs updated

seamonkey-latest2.53/source: Changing branch to 2_53_12_beta_01...

well amazingly that works

ok the final path is xr.binaryoutcast.com/seamonkey-2.53 it is just simplest for my setup

it is indexing right now

unlike orginal mxr i exclude MOST tests and sqlite3.c so it doesn't take 45-90 minutes to run the identifer search

just be aware ok frg_Away IanN

ok

IanN: must be familar seeing that working again

heh

especially for you my good and dear friend

thanks bookmarked

i have it set to pull at 0 6 12 and 18 UTC .. should be sufficient

and especially now that i HOPE i have it ONLY running the regeneration if needed

thanks for the push to actually do this

it deff helps

oh.. ident doesn't have rust support yet.. if that matters to anyone

i could add it but i'd have to know the language constructs to do it

I personally have little motivation as rust code doesn't help my process at all

oh frg_Away have you had any luck getting nss past 3.64?

a-865 they don't care. The last issue Waterfox opened was a gem. First some back and fouth between units and then basically a fork off. From a market share perspective I could understand it. But trying to give yourself the good buddy image and then doing what any shitty corp does... Well if you have no expectations I can't complain.

cause in my tree while I got it working with gyp.. it started well spewing odd and illogical errors within dom indexeddb and gfx components for no rasin

NewTobinParadigm It should work fine but I didn't have the time yet. Still on the esr78 version but now already three backports. Linux distros compile against it and as far as I know it works there. Not that I like it.

ah right.. system nss

Yes

i don't allow that .. another complication i don't need

well in-tree on my codebase i got it up to 3.64 fully functional

but no higher

nss and nspr now apply clean to our tree but I always check the bugs for needed other changes. Neither me nor IanN found the time yet.

frg_Away: did you try teaching nss to not build with avx2?

i dunno if it is required for your level but it was for mine

Didn't change anything here with the ESR78 versions. Still rune fine on Core 2 Duo. Offcial builds use VS2019 for x64 and clang 9 for x86 under Windows. clang 9 macOS and gcc 8.3.1 Linux

that was the gyp followup to code.binaryoutcast.com/projects/aur…c3249161c8eb836abda35df351609fc5102

oh

tomman: I blame google and w3c :-P

that was because msys 1

ok

you use mozillabuild 3+?

for you

mozilla build 3.4 latest and greatest

yeah the codebase requires modifications to use it

i believe

i haven't researched it yet

just been spot updating mozillabuild 2.2 with updated components

nsis 7zip and the like

but keeping msys reletively untouched

just got tired of using some versions of stuff from 2016 and older

especially nsis and 7zip

ah ident is complete

I have not build a 2.49.x in some time. Needs VS2017 and mizillabuild 3.3 or 3.2. Some of the defines are not expanded correctly in the build files. I needed to fix it for 2.53

<+tomman> it is like whoever works for KaiOS has never used a cellphone in their lifetime <-- so, like many android smartphones?

yeah but your vs2017 hack was still using the vs2015 compiler i believe

in order to move forward with my plan which i no longer have the skill pool for to update spidermonkey wholesale.. i need to first get the codebase building with 2019

So 2.49.x is confied to 3.2

if shit hadn't hit the fan by now we'd be pursing our plan to start wholesale updating spidermonkey up the ESR train

i have little doubt i could get it to ESR60 by my self.

2.53 is up to VS2022 now. Needed some changes for the directory structure. Mozilla dropped support for it but I like to have a second compiler around.

i have the build system and research skills to manage it.. but past that i am sure it would get harder

there is no point to going past 2019.. one because visual studio is fully a service

and two.. Mozilla went clang

so why bother

by the time older visual studio is hopelessly obsolete and unobtainable.. windows as a platform won't BE enough of windows to matter anyway

I am now well into 58a1 patches but it becomes harder. Need to drop legacy generators which are basically broken already and the object getters and setters are really in the way too. nsIParserService and xmlrequest also.

i don't see a reason to drop stuff just because it is the progression

VS2022 is like VS2019. I only use the compiler from Community

outside js and dom and layout things could stay reletively the same for nearly ever and no one would care

i mean am I wrong?

gfx is a problem

in what way?

what old ass cairo?

angle?

cairo was royally screwed up

upstream

and mozilla never fixed it because webrender bypasses the issues

new formats and angle harfbuzz and the other stuff.

widget too. macOS and wayland needs updates.

heh.. i don't even HAVE wayland

i dunno if i will get it.. or if it is even worth it.. as wayland emulates x11 anyway and will never be allowed not to

We neither but it is just a matter of time till the 20 year old alpha phase will make way for a 100 year pre beta phase and x11 be gone.

I don't care about http/3 and quic and all the other junk but this might be needed too later.

frg_Away: did you ever add reader to seamonkey?

reader mode

No

Thanks to the everchanging web I am now doing 90% backports. Glad that IanN still finds some time to add new stuff to suite. I only now and then when I am fed up with ports.

frg_Away: you need to sort out cloning!

IanN well you failed me here. You promised!

frg_Away: task 9999999

frg_Away: or maybe task ∞

nsIClone

that sounds like a Netscape function...

a scriptable interface to some ye

IanN task numbers starting with 99999 are reserved for me personally so better make it 99998.

IanN: did you forget xpidl?

xpidl sounds like ex-piddle!

ex pea eye dee el

cross platform interface description language

right, off to eat, bbl

enjoy IanN