hi...

frg_Away: unfortunaly, I have the same linking issue

I added: export LD=lld15

but then when I see the errors, it is still ld.bfd

is there another way of setting the linker? I remember a real flag instead of export

Sorry can't help here. I know it is possible to compile under FreeBSD but never tried and don't run it. Might just missing a path statement or libray. Maybe post your mozconfig.

I used to compile it for many versions, so it was possible.. but I did also upgrade FreeBSD

mozconfig is always the same, except for me forxing clang versions

I think tthere is a different way of setting the linker, but don't remember off-head

mozconfig: pastebin.com/7EaqPNaN

Maybe change the LD_LIBRARY_PATH.

hmm.. usually it is a bad idea in my experience

do you like my mozconfig?

grey_gandalf: I am all about mozconfig aesthetics and sanity .. yours is sane and aesthetically fine. does install-strip work.. haven't seen that directive in ages..

not a huge list of enabled system libs.. already a plus in my book

I wonder when the last time seamonkey got a successful freebsd build..

this would also build seamonkey with clang

well.. system libs are plus and cons... at the end it just saves build time

as far as I know freebsd codepaths would expect gcc i dunno if the clang codepaths have bsd special cases not-mac

it makees sense if you offer a binary package so it is consistent, less for user builds... or I risk breakage everytime I update

FreeBSD is clang-based, remember

171208 Jul 2 2021 seamonkey

wow...but this is inconsistent with my build config:

well so post-56 patchqueue that's a good sign

SeaMonkey 2.53.18.1 - 20240206085724

very recent and back then I did build with clang15, as I am trying now

can I see your mozconfig when you get a second

the .18 one

there is no evidence that systemlibs are an issue on 253 .. part of my resistance to systemlibs is the burdon when you are 3-5 people and doing windows as well.. plus at the time a rash of systemlib abi bustages were happening all around the same time soo the policy i came up with was systemlibs are mad m'kay you shouldn't compile with systemlibs and expect it to work without issue

but that doesn't apply to seamonkey which has stayed closer in this regard to mozilla and thus updated some stuff in response

that and combined with ancient firefox mozconfigs being expected to work with something that was not configured like firefox but the configuration its self had long changed before i got my hands on it.. the 2010s just sucked grey_gandalf_

the .18 is the same I use for .20 :) copied it over

except for llvm-config

grey_gandalf_: now you're gonna have to build .19 to see if that works

that is what I would do

didn't build... rust issues... that's what I waited for 2.20

I think I am unable to rebuild 18 too

downgrade system rust?

or rustup to 1.76 and to victory maybe?

one rust I have :)

mozconfig looks ok but you shouldn't need the 15 appended. Should pick the right version.

back then I used 1.74

74 is fine too

i have tested this

now I have 1.81

73 as well

freebsd doesn't offer multiple rusts anymore

1.76 seems to be the sweet spot for me anyway seems reasonably compliant and up to date

after that it has resistance again

I upgraded freebsd in the meanwhile...

there should be an obsolete 2.53.19 patch in the bug.

1.73 if you need macOS support fully retained.

oh right that

well i wonder how people would feel if GCC suddenly decided no linux kernels before 6 can run gcc whatever number they are up to

how well that would go down

greetings buc

grey_gandalf_: For issues with system rust, try Fedora's patch for 2.53.20: src.fedoraproject.org/rpms/seamonke…onkey-2.53.19-mozilla-1882209.patch

nsITobin: hi

so after scanning this patch.. the natural response is to want to rip rust out correct? I mean we can't for seamonkey but that IS a natural response correct?

hi buc

nsITobin: the patch fix the code for the rust versions >= (afaik) 1.75

frg: hi

It seems that the patch is needed for all the system rusts in all the distros nowadays (even rhel8 requires it).

But SM upstream uses rust from Fx, which is still old. Let's hope they'll update it soon too.

i didn't realize, but you cannot delete history items from the urlbar dropdown

is that expected behavior?

therube: is the urlbar still partly xpfe?

cause it has two different dropdowns depending one connected to places and one connected to its own form history

(you're talking greek to me ;-))

i don't think this is form history, but it would rather be (places) history

form history would relate to field "history" on a web page

buc I can't update rust to later tahn 1.73. Would loose partial Windows and macOS support.

therube: seamonkey's URL bar is decended from the original netscape xpfe version it only does places as a hack to its own urlbar form histroy it keeps seperately

luckily frg has been updating places and some of the FE bits should be updated once the places ui and backend settles

therube: I have the go to history menu bit in my sights and no promises but I was experimenting towards fully converting the URL bar off xpfe autocomplete.. it nearly succeeded but has some issues and I abandoned the work.. could revisit it later on too

that would put the urlbar under unifiedcomplete or at least placesautocomplete depending on what is in the tree at the current state..

which I believe may be a pre-req for really fully enabling places under the urlbar but it has been a few years since I last looked

it needs to be unified but not as Firefox has done it. This implementation s*cks imho. Task 9998

yeah can't disagree there

i'd go back to global history and bookmarks.html if it was still existant

and was my decision

i just don't understand the places architecure

not any further than the code using it in place

places is ok but not the unified toolbsar. Needs a dropdown marker and more options

oh unified toolbars nah

well you already retain placesautocomplete component xr.thereisonlyxul.org/seamonkey-2.5…rch?string=placesautocomplete&find=

yes but own impementation. hacky

frg_Away: i can't complain.. repo.palemoon.org/MoonchildProducti…ents/places/nsPlacesAutoComplete.js

I don't have a problem with rust at the moment, but a link issue with libsqlite

of all the things..

missing to link standard symbols als ceil, floor, exp and pow!

frg_Away: I am not opposed to dirty filthy hacks.. I am opposed to dirty filthy hacks being called solutions and not dirty filthy hacks that should be replaced as soon as some better code comes along..

HTML5 is one big filthy hack on Mozilla I discovered by following its initalization in dom.

and they don't seem to have an issue with that

so I do ;)

also unintentionally showed me how I can just tack on any thing in its place should something show up.. which is kinda neat i guess but only for codebases at seamonkey or below..

I have a clue to the issue

why do people thing Linux = Unix? :)

UNIX_BUT_NOT_MAC

XP_LINUX is a subset of XP_UNIX

frg_Away: if you have a moment would you do a quick check to see if this works with your ftp creds cci-dev.thereisonlyxul.org/panel .. the logout uses a javascript trick to invalidate the basic auth so js needs to be enabled.

frg_Away: I think I found the fix. ALthough I have a doubt the build system works fine

should I open a bug report for a build system change?

what is the potental fix?

in third_party/sqlite3/src/moz.build

it first enables SQLITE_ENABLE_MATH_FUNCTIONS

then adds library only for LInux

I changed the line to if CONFIG["OS_TARGET"] in ('Linux', 'FreeBSD'):

seems reasonable would this be from a mozbug or a mozoversignt needing a patch

of course, maybe it is needed also on OpenBSD or FreeBSD or Mac... who knows

well i don't see it breaking things especially if you say it fixes it and it won't affect other target platforms

just test

if seamonkey was more static I'd go through and re-evaluate the XP_* conditions resolving to as common as possible stuff that needs to be linux specific would once again be linux specific and unix common code would be better marked for the kind of cross-platform the codebase was supposed to have not cross-major-platforms

for some reason I read that as "Maxtor Platform" .D

and then I would say "Maxtor dude!? If you want to reliably lose your data, get a Seagate"

that was in 2010 when my maxtor drive failed for the last time

lol

besides, as a Captain in a fictional future organization .. i outrank an army major.. so THERE not EVEN in my chain of command.

btw I figured out why Elon likes crypto so much .. he is taking the phrase cryptofascist literally..

if we want to talk about fictional future organizations

I will make sure LCARS is based on XUL and KaiRo's seamonkey skin..

still seamonkey doesn't start for me with "mach run" and don't want t ruin my existing installation my installing it

can it package its self

nsITobin I can login

thanks frg_Away always good to make sure stuff works for one than one person

go to dist/bin and start seamonkey directly. If it does not run something is wrong with the binary

back up your profile

shit... I cannot file a bug using seamonkey itself :) need to startup firefox

./mach run merely issues no remote and a specific place for a temp profile anyway

grey_gandalf_ Mozilla does not have the freeBSD fix in mozilla-central. I can add it when you get the version working

well, with my fix build finishes

I hope it is smart enough to resume abuild, link and finish :)

when I run, I get

ModuleNotFoundError: No module named '_sqlite3'

still something wromng with sqlite3 it seems

indeed, but what? I wonder if the build is complete or not.. maybe I need to clobber and rebuild for safety

always clobber when you change something binarywise

just to be sure

least until you get a feel for when and when that won't matter

err that phrasing could use work LOL

least until you get a feel for when that will and won't matter

there we go

and even then clobber anyway sometimes

shouldn't be... I jsut changed linking, moz should finish linkgin and continue

it wasn't so unreliable, with the older version of ArcticFox it still does quite wqell.

I will clobber though. But runs for quite some time here.. maybe better close apps and save memory

however there is a matching bug report made by myself :)

howver since I use the same computer as 2 years ago... all should be fine now, ******

Did I ever metnion I hate python second only to rust?

the answer is py311-sqlite3-3.11.10_7 vs py39-sqlite3-3.9.20_7

multiple pythons....

grey_gandalf_: are you amazed that other mozilla rebels might agree with you?

no, and I hope beyond mozilla rebels too

python is a nightmare

it has nearly ruined mainstream php

composer and frameworks .. php IS a damned c framework for the web..

rust is pythonic javascript cosplaying as C++

to me

when I learned rust has no dynamic shared libraries... I thought, are they nuts?

disconnecting and installing!!! hope to see you sooon

frg_Away: if you can read me... know that I am typing from IRC running in 2.53.20 happily compiled and installed

spectacular

grey_gandalf_: we need more of that

I seriously believe if it is manuvered correctly and there are plenty of options it MIGHT be possible to squeeze out a segment of the market passionate about keeping an actually open World Wide Web in existance with some kind of first party support.. This udp binary app stream web won't solve all the problems and the more they force it the more alternatives should spring up or become more viable in the minds of people.

and never forget the potental of .. well these guys think there is still something to it..

I do however also believe this is likely the last stretch and chance to do that though before its technologically impossible to keep playing

so to speak

how complicated has it become to simply run a website when not using a one click service?

kinda ridiculously

and this is where living standards break down.. people can barely keep the computers up to date for it.. people is even harder them selves..

nsITobin: more of what?

browsing a little... opening mail... seems to work in quick tests. Cool

frg_Away: only two patches nneded: your rust-fix for more modern versions and the sqlite3 fix

could have been worse :)

I will put the fix into 2.53.21

more happily compiled builds of not-major browsers

especially seamonkey

wonderful...

grey_gandalf_: Error 418

what's that?

sorry, hacking on Gnustep and ArcitcFox at the same time :)

look it up

it's in an rfc

rest well the frg

oh, no teapot... but why do you reference this, while I was happy to have an up-to-date SM ?

right now, I'm more with spirits than with coffee

because it is fun

and fun things need to be mentioned and remembered

for fun