just installed seamonkey

on el8

was a startup url

oops

yeah that will deff work

and vmware installed.. no hassle just works

it's amazing what can be achieved with the correct technology

everything is better because it is EL8 and /at most/ Windows 10 LTSC 2019 across everything now.. I just don't think I am gonna use post-pandemic operating systems anymore.. Over the past few years it simply hasn't gone well.

VMWare is dying, yo~

friends don't let friends get Broadcom'd

tomman: it is dying for legacy VM needs like anything pre-vista and soon pre-10 i am sure.. but for my purposes it shall serve.

its just a Windows runner and then only a windows runner to provide me decent access to my 2005 era XP programs and building the seamonkey without rebooting

I used to be a VirtualBox fanboy, then Orrible® ruined the fun (I can't even access its website anymore, much less legally use most of it because Das Sanctions™)

seamonkey i could build with kvm

but I still NEED a fully windows environment for.. Paint Shop Pro

and never been a fan of Hy-Perv, er, Hyper-V

games run on proton.. notepad++ and winamp run on wine

To be fair, I haven't tested PSP on Wine lately

cause it continues to whip the llama's ass despite the new rights owners

tomman: it can't be installed

I still use PSP 7 on my old vintage boxes, runs like a champ on everything from 98SE upwards

msi installer issues

i need exactly PSP 11

oof, anything post-Corel doesn't exist for me

only PSP11

7 was the final Jasc-era version... or was it 8? But 8 was meeeeh

is sufficient

ONLY that version

not 10 not 12 not 2021

its the ONLY corel version on par with jasc psp9

i COULD use that but my font rendering and some extra features would be sacrafised

and none of my stuff would be readable

i am locked into PSP11 or newer and that means JUST 11

tomman: I would need to do a staged install on windows capturing EXACT regkeys and files then overlay it to wine for even a CHANCE to get it running

and frankly I have no facility for the reg part

Let me guess: you have tried even on recent wine-devel 9.x versions and met with failure?

MSI installers are indeed a pain, sometimes even under Windows itself

the corel msi installer has embedded IE control i think THIS is why it will never work

MSI was a nice electoral promise, but I've seen it break BADLY on Windows itself, and on Wine it's simply a world of hurt

nah, MSI doesn't use embedded ActiveX controls

MSI is just RPM for windows

tomman: it literally uses IE to display a webform for corel signin

its a custom msi installer

it even clicks

with the IE navi click

on windows

if the sound is set

oof, I forgot those crazy mess were possible, but very few even bothered because it's horrible to author installers like that

but then this is Corel we're talking about, so...

Yes.

lol

PSP is really the only thing requring a graphically accelerated UI for sheer usability sake

well i'd say get the versions you want softwareupdate.vmware.com/cds/vmw-desktop

cause i doubt that is gonna remain free and clear forever despite them making workstation free for everyone for personal use

tomman: graphic design wise I'd be fine with an XP64 VM

but that is heading into needing virtualbox to get any kind of decent performance from that era of OS.. and i can't have more than one VM solution installed else it kills the bus

or rather the network stack the buses need

On windows least windows today loss of the network stack momentarily or at all isn't fatal in linux.. it is

frg_Away: running the patch queue now

frg_Away: let me know if you are interested in updated hg and tortoiseHg Workbench.. I just built fedora 34's version.. gonna later push it to whatever last qt5 version is

for el8

hell I should create a repo

nsITobin hi. I picked the stock one. i think 6.63 currently

there isn't one for el8 in epel or rocky sources

unless you are using a snap thing

Pretty sure I use an rpm. Let me check

if there was one it isn't in epel anymore

its deps are

flatpak?

the QT5 pyqt ui looks and seems to work better than the qt6 one.. I don't think there was a lot of testing

hi buc

Hi frg

tomman: if sanctions cause problems with Oracle downloads, how are things there for linux distros, is that at least a way to get such packages?

frg_Away: configure is happy and mach create-mach-environment was smooth and my exporting state path and my reduced artifacts are all working

post-Corel? yeah, the only brand I had associated with paint shop pro in my memory was jasc :-D

who owns it now?

corel

nsITobin I use the centos8 mercurial repo. So far it works

there is one?

www.mercurial-scm.org/releases

frg_Away: yeah kernel 4.whatever does NOT handle IO priority as well as newer kernels..

i am gonna have to build a newer one

ah i see

neat

all i saw on the website there was just go get it from a flatpak or snap or whatever or from source if not fedora debian

nsITobin: for el8: elrepo.org/linux/kernel/el8/x86_64/RPMS

yeah tried one of those

video didn't work

likely did it wrong

I'll just str8 compile a fat kernel

no initramfs needed

ah yes, as reported in the newsgroups, MICROS~1's deadline for forcing OAuth2 is not really a deadline, but probably the start of a transition. Or something.

just leave kernel modules for vmware use

I can still access @hotmail.com with regular auth, after the first failure, which, from another post there, seems to be actually intended

njsg: and this is why Interlink Mail & News was discontinued that and attacks from MCP personnel

I still fail to see what is so secure about oauth2. Just looks like another vendor lockin to me. Not 2 SFA

vendor lockin AND ties an unrelated internet protocol to the initiatives of the openweb as dicted by whatwg

I'm not sure it adds that much to the concept of "app passwords"

that one already has different passwords for each application (although this then might depend on the user not reusing for more than one); these could be made time-limited to achieve a similar effect as OAuth2, if the concern really is security

kind of looks like designed in the same vein as web-based 3DSecure validation, where you're asked to interact, to confirm the transaction, via the site you want to do the transaction on

OAuth2 opens a browser window triggered by the client. I'd personally consider perhaps having the user go separately to the account control panel generate the password

but, other than this all? It adds a huge complication to authentication

and now provides an easy way to prevent third-party clients, which usually wasn't a problem so far with regular authentication

because, at least with Microsoft, they do check the program ID and they have a set of allowed clients

this is basically paving the way for Hotmail to become what it was in the late 90s/early 2000s, a service without an open protocol, that you could use only in, say, Outlook Express

yeah and then you will need an app password AND a code from your phone

and your normal pasword

and a code from the email address you are trying to access

njsg: tell me I'm wrong lol

nsITobin: I once achieved the magnificent scenario of a recursive e-mail validation loop. thankfully, one of them wasn't mandatory, otherwise I'd have been locked out

mattatobin.com expiring makes it so i can't get into my email at namecheap to get the code to get into namecheap to pay name cheap to either renew my domain or renew the email service

njsg: better than an unpatched 20 year old xml recursion OOM bug

because no one could agree on a depth

did mozilla ever decide one one.. or did they solve the problem by pushing xml into near obscurity

better than java at least back when I was learning it in university

java.lang.NullPointerException is my experience with java on everything

and we were told that Java lacked tail call optimization

I like java sorry :)

frg_Away: elfhack works

ac_add_options --enable-elf-hack

ac_add_options --with-linker="bfd"

export LDFLAGS="-fuse-ld=bfd,-Wl,--no-rosegment"

from what I understood back then it was a thing with java specifically, and the language has meanwhile added support for that IIRC

the JVM either always had, or at least had it already back when I started using java

export LDFLAGS="-fuse-ld=bfd -Wl,--no-rosegment"

ok after some further testing it seems the with-linker option is incompatible with LDFLAGS that depend on specific linker options cause the fUse is being placed after.. causing a configure test error .. we need to get configure to put its with-linker value before adding any LDFLAGS from the env

until that is done JUST do export LDFLAGS="-fuse-ld=bfd,-Wl,--no-rosegment" with --enable-elf-hack

really --enable-elf-hack should be setting LD and flags something to enhance later

but as I said to buc figured out what to do to work it as-is.. tested the release queue and the updated build system enhancements which do make things better and have an idea what to change/enhance later to actually fix it

all in all successes across the board for the me.

thanks to all of you

no scratch that it still doesn't work

why did it work before

ugh

ugh

because you need to have both the configure flag and the ldlags else the build system isn't aware of it

frg_Away: I think just the configure argument to set the linker to bfd is all that is needed the LDFLAGS seem incompatible with it and only in my original incorrect form were they silently ignored otherwise it errors out

and those flags may be needed for newer fedora possibly but if this build works and it is reduced it will mean just need --with-linker=bfd will know shortly

frg_Away: ok.. final instructions least on EL8.. --with-linker="bfd" --enable-elf-hack and that's it

so njsg Kernel: arch/x86/boot/bzImage is ready (#1) any bets if it will boot or not?

man that is SO much better.. I should still -j10 or -j8 but at all cpus nearly pegged I am only getting the very occasional stutter from winamp and x11 .. this was virtually eliminated in the 6x series but this is .. acceptable.

of course now I am responsible for rebuilding it for updates and if i didn't have a feature enabled.. but yeah no initrd builtin gpu drivers.. some minor config.. easy everyone should do it at the slightest bit of resistance from what is handed to you

least now the kernel is efficient enough to unchecked building AND listen to music

Linux 5.15.167 #1 SMP Wed Sep 18 14:33:57 CDT 2024

according to seamonkey

I'm trying to find information about serial mouse protocols. For some reason Google gives much better results when I do the search using Firefox than when I do it using Seamonkey...

The results that Google gives to Seamonkey don't even make any sense anymore

The results for Firefox are already bad and don't contain the information that I need, but the results for Seamonkey are even worse

electron/electron #15691#issuecomment-2161904711 Word of God™ says "any platform other than x86-64 and ARM shall not exist for us, don't bother sending in patches"

try startpage

google results are constructed based on your browsing habits and previous search history and account data

logged out google is useless

logged in its still 60% useful cause it has a decade's worth of searching about mozilla shit

:P

google empowered mozilla to ruin everything.. google empowers me to undo it

besides google isn't really as high on the list as it was still top five but microsoft is

tomman: so they are ignoring RISC V

they're definitely ignoring PPC, for sure~

it's curious that indeed there is still no RISC-V Android stuff

but then the current RISC-V SoCs are meh at best, and Android is a bloaty beast

so who wants to help me fork a rust-free kernel?

the Western Digital fork of version 5.0?

I heard WD loves RISC-V