serpente has quit [Remote host closed the connection]
SiFuh_ has joined #crux-social
SiFuh has quit [Ping timeout: 272 seconds]
<SiFuh_>
remiliascarlet: I only care to see the outcome of his sentence in Korea. I don't even want to look at his ugly face.
<remiliascarlet>
2 weeks ago, Codeberg was so-called having a "spam attack" of somebody tagging everyone in an issue, and write "[truth/truth] NIGGER BALLS", and then they wrote an official statement specifically alienating everyone who's right wing, conservative, libertarian, centrist, and anyone else on the right of Stalin. As time went on, it became clear that it was all an inside job. Last week there was a
<remiliascarlet>
so-called "spam attack" in Gnome's GitLab repo in a very similar way of somebody using a certain person's name they hate (Lunduke), which was super obvious that it was an inside job. And now you have the developers of Asahi Linux both quitting at the same time to become a transformer VTuber, from which we learned that they're most likely the same person, is a transformer, is from Spain, and
<remiliascarlet>
lives in Tokyo.
<remiliascarlet>
What all 3 of them have in common is that they all claim to be "attacked" by "far-right wing forces".
<remiliascarlet>
And all of them are most likely just inflicted by themselves.
<farkuhar>
Why does zathura (a document viewer) need to link to libseccomp? zathura would look out of place if added to the list $(grep -lrE '^# (Depends on|Optional):.* libseccomp' /usr/ports/), but apparently it does belong there.
<farkuhar>
And by "need to link" I mean that zathura eagerly links libseccomp if the latter is found during build, not that it fails to build if libseccomp isn't installed.
<SiFuh_>
farkuhar: zorz: ukky: https://i.snipboard.io/OqVCAd.jpg Can now watch the gravity when fermenting beer in real time from my office without the need of a cloud server located somewhere on the planet that I have no access to if I lost internet;-)
<SiFuh_>
You measure the sugar/dextrose content in your beer wort as it ferments. OG is the original gravity. SG is the specific gravity (As in what is the reading now. FG is the final gravity but this device doesn't have FG. The device floats in the beer wort and tilts as the sugar gets consumed and transmits a calibrated reading.
<ukky>
SiFuh_: It looks promising, but I have no idea what I'm looking at.
<zorz>
again joo joos
<SiFuh_>
So for example this one has an OG of 1.046. Water with no sugar is 1.000 but because of my altitude and the weather it reads as 0.994.
<zorz>
SiFuh_: the beer maker....
<SiFuh_>
ukky: zorz: The old one died. Corroded terminals. But I hated how it transmits to a cloud server somewhere. I wanted something localized so I contacted a guy in Croatia to send me this.
<SiFuh_>
Also it is modular and opensource which means I can modify or repair as schematics are available online.
<SiFuh_>
ukky: zorz: So my average beer I make has an OG of 0.023. That tells me the density of the wort has a reasonably high content of sugar. When the yeast consumes the sugar the density changes. So the device tilts, it's angle changes as it floats, until fermentations tops. Then after 2 days if the gravity hasn't changed and is less than 1.010 it can be bottled. In my case 1.004. This is good because you
<SiFuh_>
won't have exploding bottles with left over sugar in it.
<zorz>
SiFuh_: this is greek to me:p
<SiFuh_>
Yes but knowing the OG and the FG you can calculate the ABV of your beer.
<SiFuh_>
ABV = Alcohol by volume.
<ukky>
zorz: So, which motherboard did you end up buying for your project?
<ukky>
zorz: I have to allow so many scripts on minisforumpc.eu just to see the image of that MB
<ukky>
zorz: AtomMan G7 Pt is too smOll for me.
<ukky>
zorz: but if you are restricted in space, then NUC-form-factor is okay I guess.
<SiFuh_>
Mine mini PC can support 2 x NVMe and 1x SATA
<zorz>
the new nvme 5pci need coolers, things get complicated.
<zorz>
heh
<zorz>
no i have space...
<SiFuh_>
I have 1TB NVME another 1TB NVME and a 2TB SATA drive
<farkuhar>
Huh, the bootstrap.py that ships with rust has a hard-coded fallback value for the ABI: unknown-linux-gnu (line 336, in the case 'elif kernel == "Linux":'). For CRUX-MUSL we should probably add a sed command to correct that omission.
<SiFuh_>
Interesting. I mounted an NFS filesystem on Winshit. Disconnected from the lan. Modified a file on that NFS server. Saved it. Then realised "Fsck... I am not on the lan". Reconnected to the lan and file had been saved. Must be Winshit Swapping or RAM
<farkuhar>
When building rust 1.85.0 I didn't notice that bootstrap.py made no allowance for the ABI unknown-linux-musl, maybe because I already had a working older version of rust and didn't try to run a full bootstrap.
<ukky>
Most of my drives are spinning drives. Only NUCs have NVMe primary. But for testing I use SATA SSD, they are cheaper than spinning drives if you need plenty.
<SiFuh_>
ukky: My SAMSUNG SATA 8TB SSDs are pretty good. No issues yet in 5 years
<ukky>
Yeah, Samsung EVO SSD are not bad, I use two EVOs at work.
<ukky>
My 11 y.o. MSI motherboard has two BIOS flashes. Nice feature to have. But DDR latches suck on that MB, MSI has latch only on one side of the DIMM. And no serial port, that sucks too.
<zorz>
my laptop is msi, almost same as sifuh. sometimes i had it working 30+ hours at 95c. still works.
<ukky>
95°C is a lot. Koodos for MSI if they can work in that condition.
<zorz>
hashcat, and when i burn substitles
<zorz>
ok subtitles usually is for 20-45 minutes
<SiFuh_>
ukky: It's true, it runs freaking hot.
<zorz>
ukky for this laptop they say critical is 110c
<SiFuh_>
Mini PC's usually run around 55 to 65 unless I am doing video processing.
<SiFuh_>
But the laptop is stupidly hot. Rough the right speaker and it is enough to cook an egg off of.
<zorz>
hahahaha
<SiFuh_>
Rough/Touch
<zorz>
burn the table
<SiFuh_>
zorz: I can handle quite hot temperatures. But sometimes it runs so hot that I can't place my hand on it for very long.
<SiFuh_>
zorz: Reminds me of the guy who burned his penis
<zorz>
right now [zorz@void ~]$ grep -o '[0-9]*' /sys/class/thermal/thermal_zone0/temp | awk '{printf "%.1f°C", $1/1000}'
<zorz>
92.0°C[zorz@void ~]$
<SiFuh_>
ukky: zorz: I poured a shot of vodka when I was doing the graphing and calibrations. Then noticed the shot glass was empty. Thought, may as well pour another. So I did. Then realised I had two shot glasses. One on the table and one the cupboard. Hehehe
<zorz>
oooo gggggg
<zorz>
if only they had better bios. msi is cool
<SiFuh_>
It's a pain in the ass doing the graphs. Need to use water a hydrometer, thermometer and the device. I have to get two readings for gravity and two temperature readings and one angle reading. Then plot it in. Then add 1/2 a cup of sugar and do it again, until I use 1 KG of sugar. It took 4 hours to get 10 lines to plot. The device needs to stop moving in the water. You need to take multiple readings
<SiFuh_>
to get an average.
<ukky>
zorz: you mean, BIOS settings?
<SiFuh_>
ukky: There is a code to unlock the bios in the MSI to get many many more options. But when I unlock it. The laptop will not boot again.
<zorz>
i mean this one
<zorz>
[ 9784.679853] ACPI Error: No handler for Region [VRTC] (000000001db14ada) [SystemCMOS] (20240827/evregion-131)
<zorz>
[ 9784.679866] ACPI Error: Region SystemCMOS (ID=5) has no handler (20240827/exfldio-261)
<zorz>
[ 9784.679873] ACPI Error: Aborting method \_SB.PCI0.SBRG.EC._Q9A due to previous error (AE_NOT_EXIST) (20240827/psparse-529)
<zorz>
[ 9983.698692] usb 4-2: USB disconnect, device number 3
<SiFuh_>
zorz: Mine is similar
<zorz>
the only issue in dmesg
<ukky>
SiFuh_: confirming your words, I have seen 'unlock' with regard to MSI BIOS.
<SiFuh_>
ukky: We have no legacy mode. So I unlocked the bios with a cheat code and selected Legacy and then it wouldn't boot back into BIOS.
<zorz>
thats why i say, next one gigabyte.... and for monitor gigabyte
<ukky>
zorz: Those ACPI errors are definitely a bug, but OS can live with them.
<zorz>
yesssssss no issues
<SiFuh_>
I am going thinkpad if I ever go for a new laptop
<ukky>
SiFuh_: 'back into BIOS', like you cannot enter BIOS settings, or system does not boot?
<SiFuh_>
ukky: Both
<SiFuh_>
It turns on and sits their with a black screen
<SiFuh_>
there*
<SiFuh_>
ukky: The backlight is on but just looks like a glowing egg plant sitting on the table doing nothing
<zorz>
hahahaha
<SiFuh_>
zorz: Don't rewind. Check out this. Hehe. not bad how slowly it went down and how it just came back up again. https://youtu.be/CDg9E0NqlYY?t=1992
<SiFuh_>
ukky: zorz: That Jeram Janggut trip with 25 off-road vehicles that left everyone including me with damage and breakage, I had heard and seen many videos of trucks being rolled and it is apparently very common. Fortunately none of us rolled.
<SiFuh_>
farkuhar: ukky: zorz: https://i.snipboard.io/ngcPNQ.jpg This is the graphing I was talking about what I had to do to calibrate the device.
<zorz>
in greece they start placing stickers on tesla cars, I bought it before Elon lost his mind.... hahhaha dont hurt my car:P
<SiFuh_>
Heh
<SiFuh_>
Idiots
<zorz>
indeed funny
<ukky>
SiFuh_: What do you do for fun? I plot graphs in my spare time.
<zorz>
ukky: in my spare time for fun, I drive mad sifuh
<SiFuh_>
ukky: No I had to. So that I can get the three lines below, or the device I bought from Croatia would be useless for monitoring my beer fermentation.
<ukky>
zorz: you know which button to press
<SiFuh_>
ukky: I didn't want to do it at all. Wished they were calibrated already. But as I mentioned before, altitude plays a roll.
<ukky>
SiFuh_: To do it right, of course you have to know all the details.
<zorz>
ukky: button & timing :p
<SiFuh_>
ukky: Actually the pressure sensor seems to not be functioning. I will investigate later.
<SiFuh_>
ukky: I found a Vintage Super Maid Folding Hinged Pan for 44 USD and the postage is good too. I ordered already
<SiFuh_>
I put it in my watch list hoping they'd discount the price and they did.
<ukky>
SiFuh_: Congrats. I know the feeling when price drops and you get your item.
<SiFuh_>
Yeah, usually takes 1 to 2 days for them to offer a cheaper price. This one took almost 2 weeks
<SiFuh_>
Okay, pressure sensor works.
<SiFuh_>
Not that I need it. I don't ferment under pressure.
<SiFuh_>
I find that fermenting under pressure, I don't get Lactobacillus, but I like beer with a hint of Lactoacillus sourness in it for the extra flavor
<SiFuh_>
Oh crap, I still got vodka in a shot glass.
<SiFuh_>
farkuhar: In the US when you pay bail, do you get it back if you turn up to court?
<SiFuh_>
Bail in the United States refers to the practice of releasing suspects from custody before their hearing, on payment of bail, which is money or pledge of property to the court which may be refunded if suspects return to court for their trial. Bail practices in the United States vary from state to state. -- Wikipedia
<farkuhar>
SiFuh_: Interesting question. I can't remember how my state handles bail; I'd have to look it up.
<SiFuh_>
zorz: The reason I stopped using void was because X-Plane 11 was a pain in the arse to set up on it and then when you finally had it running the audio refused to work. That was 6 years ago. I decided to install void and X-Plane 11 a couple of days ago, and it still has the same fscking problems. Works fine under CRUX, ParrtOS and Slackware.
<farkuhar>
SiFuh_: Did you open a bug report with the Void devs?
<SiFuh_>
farkuhar: In Australia we have two bail systems. One through the courts and another through the police. The courts will only ask for bail money if they think you will run and not appear in court. The fees are not very high though. Couple of thousand dollars. But the police bail is strange.
<SiFuh_>
farkuhar: No, because they are all wankers.
<farkuhar>
Even if they're all wankers, they still deserve to know that their audio setup is broken with X-Plane 11.
<SiFuh_>
Police will arrest you and rather than keeping you in a cell overnight they can send you home but you should pay bail. But the thing is that it is so relaxed that just to keep the books tidy, they usually ask for 5 cents.
<SiFuh_>
farkuhar: The audio I am referring to uses that other one. I can't remember the name. I think it starts with A.
<SiFuh_>
OpenAL
<SiFuh_>
I can call you Betty, and Betty, when you call me, you can call me OpenAl
<SiFuh_>
farkuhar: I know how to set up X-Plane 11 on Linux. Void just doesn't seem to function well with it.
<SiFuh_>
I will just end up install CRUX anyway. It's way better than void.
<farkuhar>
It's amusing to see how many Void packages are marked "Orphaned" in the maintainer line.
<SiFuh_>
If zorz was a package, I am sure he'd be orphaned
<SiFuh_>
Hahahaha
<farkuhar>
There's a difference between *unmaintained* and *unmaintainable*.
<SiFuh_>
Hehe
<farkuhar>
Which CRUX will SiFuh_ install ... glibc or musl? 3.7 or 3.8? Too many choices at the present time.
<farkuhar>
Hence my suggestion that ppetrov^ propose a moratorium on further commits to the 3.7 branch. Push everybody to adopt 3.8 early, even if it hasn't been officially released, and then the bugs will be discovered sooner.
<SiFuh_>
farkuhar: I don't think X-Plane will work under MUSL
<farkuhar>
SiFuh_: Why not? Is X-Plane distributed only in binary form, or can you build it from source?
<SiFuh_>
Binary dude
<farkuhar>
Cessation of commits to the CRUX 2.2 branches didn't stop brian|lfs from trying to resurrect that historic OS in a virtual machine. Someone could take similar risks with an unmaintained CRUX 3.7, if they didn't want to be an early adopter of 3.8.
<farkuhar>
"Binary dude" <- Someone should steal this for a movie title.
<SiFuh_>
farkuhar: "Dude, where's my car?"
<SiFuh_>
And then.....
<zorz>
SiFuh_: moment i play with tor
<SiFuh_>
zorz: I hope that isn't a name for a certain organ you have compressed in those tight jeans of yours.
<farkuhar>
I thought it expanded to "The Onion Router", but maybe it has other meanings.
<farkuhar>
In beerman's tongue, "Tor" has the primary meaning of "fool", but it's also used for "goal" in sports like football.
<SiFuh_>
farkuhar: You mean like gift in beerman's tongue is poison?
<farkuhar>
Hahaha
<zorz>
yyeee i need first to run tor, give some time and then change the nftables for transparent proxy. mission accomplished.
<SiFuh_>
zorz: Don't you need to play with your onions as well?
<SiFuh_>
In 2002, I was told to put this on my balls and legs because of an infection from chaffing. I think I screamed for 30 minutes. The next morning the skin peeled off in one go. It was like a perfect mould of my balls and thighs.
<SiFuh_>
Hahahaha
<zorz>
farkuhar: they mark orphaned the packages built by the team
<zorz>
SiFuh_: go give a moratorium in #crux
<zorz>
farkuhar: => Registering new packages to /host/binpkgs
<SiFuh_>
zorz: Shut up zorz. Read before you speak
<SiFuh_>
And what the fsck is a moratorium ?
<farkuhar>
The secondary meaning of moratorium is "suspension of activity".
<zorz>
hahaha
<zorz>
moratorium is what farkuhar suggests to ppetrov
<zorz>
SiFuh_: i have an idea, join #voidlinux and ask for help muahahaha
<zorz>
moment be back
zorz has quit [Quit: leaving]
zorz has joined #crux-social
<farkuhar>
ppetrov^ already declared his intention to stop updating his personal repos for CRUX 3.7, I merely suggested that he mention the idea in #crux or #crux-devel for wider adoption.
<zorz>
and i said,,, only SiFuh_ is professional in proposing moratoriums
<SiFuh_>
farkuhar: So why don't you offer your gift to #crux, #crux-devel?
<zorz>
f4ck germans
<SiFuh_>
zorz: I have a feeling that zorz's translation of moratoriums is not what he thinks it is
<zorz>
most probably
<zorz>
:P
<zorz>
this loosers https://ipleak.net/ they have DNS detection - Under maintenance, retry later. for 1 month now
<farkuhar>
SiFuh_: I think the suggestion would meet with too much resistance. Cessation of the automatic merges in the remote-tracking branch would entail almost twice as much work. Most maintainers would prefer to commit only once, rather than in two separate branches.
<farkuhar>
TBH, I'm rather curious how the automatic merges aren't causing more problems with python3-* ports, considering the different footprints they generate on 3.7 versus 3.8. Is there a post-commit hook that fixes the footprints and signatures whenever one of the python3-* ports gets updated in 3.7, before an automatic merge can occur?
<SiFuh_>
farkuhar: I suggested a long long time ago that it should be a floating variable like kernel version
<SiFuh_>
And jaeger agreed but it never happened
<farkuhar>
Well, it's not as if the port maintainer is barred from redefining make_footprint(). Here's my work-in-progress for rust-bin: https://dpaste.com/HH9Q8WPMV
<SiFuh_>
I don't like or use rust
<farkuhar>
What I'd eventually like to add is some source integrity verification. Currently, with an empty source array, the check_signature() and make_signature() functions do not consider the upstream tarball at all.
<SiFuh_>
farkuhar: At least I can read it, unlike zorz's garbage scripts
<farkuhar>
I wish I could avoid rust too, but in order to maintain firefox built from source (the only way to enable the sndio backend on Linux), I have to put up with rust.
<SiFuh_>
I would like to see it only as a make depend and not a run depend.
<farkuhar>
Unfortunately, as a source-based distro, CRUX decided not to make such distinctions in the dependency line. But at least `prt-get depinst firefox-bin` does not trigger the installation of rust.
<SiFuh_>
farkuhar: It can always be built inside of the port and then not installed
<farkuhar>
It would be an incredible waste of CPU time to always rebuild a make-time dependency like rust whenever you needed to update firefox. Better to build and install rust just once, and then keep using it as long as a newer version isn't required.
<SiFuh_>
Ahh but here is the cool part. You can have that rust archived as well and shoved into the distfiles directory.
<farkuhar>
How is "shoved into the distfiles directory" fundamentally different from "installed under standard prefixes like /usr/bin and /usr/lib"? Of course in the latter case, the rustc and cargo binaries will be discoverable in $PATH, so you might opt for the first method to prevent eager use of the optional rust code in some other project. But you're occupying the same amount of disk space either way.
<SiFuh_>
You don't use distfiles do you farkuhar?
<SiFuh_>
PKGMK_SOURCE_DIR="/usr/ports/distfiles"
<SiFuh_>
And I meant this, not distfiles :-P PKGMK_PACKAGE_DIR="/usr/ports/packages"
<farkuhar>
SiFuh_: Yes, I do have centralized directories for downloaded sources and packages. But you still occupy the same amount of disk space, whether the built rust toolchain is installed under standard prefixes or in PKGMK_PACKAGE_DIR.
<SiFuh_>
You build both rust and firefox. Rust will be in the work directory. You technically make two packages, rust and firefox. Firefox gets tarballed and installed and rust gets tarballed and shoved in the correct directory for future builds
<SiFuh_>
rust doesn't get install under / it gets installed under the work directory for firefox. But you can still tarball up rust and save it for future firefox builds
<farkuhar>
This port-within-a-port idea is straying too far from the KISS principle, as I understand it.
<SiFuh_>
farkuhar: Tell that to Romster and Beerman then
<SiFuh_>
They all have done it
<farkuhar>
No need to tell them again, Anton Vorontsov already expressed the idea more eloquently in 2008 (although he was objecting to heidi's implementation of Gentoo-style USE flags, not the port-within-a-port practice).
<SiFuh_>
Romster is the one that told me to do it in a port I built back in 2019.
<SiFuh_>
Then he showed a Beerman example. I think it was Libreoffice with Poppler
<SiFuh_>
If I remember correctly a specific version of poppler was needed and it would not compile against the system version
<farkuhar>
SiFuh_: That's why I created poppler-ink and listed it instead of poppler as the dependency for inkscape. Even as poppler receives update after update, inkscape continues to work without recompilation, because it looks under a non-standard directory for its poppler libraries. Eventually inkscape will need a version bump, and at that time I can update poppler-ink too.
<farkuhar>
But in the meantime, who wants to be recompiling inkscape with every new poppler release?
<SiFuh_>
I don't see the point of recompiling everything that gets updated
<SiFuh_>
It should be more like a monthly thing not a daily thing
<SiFuh_>
Unless serious security issues
<zorz>
SiFuh_: go to #crux
<farkuhar>
The make_footprint() in my rust-bin Pkgfile is still missing some needed logic, because the tarballs for glibc and musl differ in more than just the random hashes. Additional filenames will need to be pruned from the footprint to avoid pkgmk MISSING mismatches.
<SiFuh_>
zorz: Why would I go to #crux when I am already there you clown
<zorz>
the conversation man
<farkuhar>
Notably, upstream builds for the musl ABI are configured with crt-static=false, just as we do for rust built from source. So we need to ensure that those files don't appear in the footprint generated when pkgmk is run on CRUX-glibc.
<SiFuh_>
zorz: There is no conversation in #crux you clown
<farkuhar>
Another problem with my flexible rust-bin Pkgfile: an empty source array, and the resulting omission of sha256 sums in the signature, renders the downloaded tarballs eligible for deletion by prtwash or oldfiles. A user who regularly cleans out the distfiles dir or the ports tree will then have to re-download the upstream tarballs.
<SiFuh_>
Doesn't prtwash have a config file?
<SiFuh_>
Secondly that is an addition to CRUX so that should be left to the user to decide how to handle it
<farkuhar>
As for what counts as "an addition to CRUX", you could say that accommodating two different libc providers in a single port is already a violation of KISS, and the simpler solution would be an overlay directory that masks the canonical (glibc) port.
<SiFuh_>
I doubt that would work
<SiFuh_>
Anyway, I am busy thinking. I am trying to remember what I am bring back to Malaysia with me if I go to Australia.
<SiFuh_>
And I need to find a safe way to dump my radioative elements without being classed as a terrorist
<farkuhar>
Heh, if you're predicting breakage from a duplicate rust-bin in the muslcrux overlay (which runs the same commands but with a different upstream tarball), then you'd be right. The resulting rust-bin was able to handle some simple projects, but for complicated codebases like firefox it doesn't play nicely with the clang and llvm compiled from source.
<SiFuh_>
But I am always right ;-)
<SiFuh_>
You can dispose of radioactive material without seeking approval if the concentration or activity of a radionuclide is not more than the maximum calculated using section 9, 10, 11 and 12 of the Radiation Safety Regulation 2021.
<SiFuh_>
A = C/MC
<farkuhar>
Anyway, the last time I hit the error "libclang: Dynamic loading not supported" during a firefox build, it was fixed by rebuilding llvm. But now I hit the same error with firefox 136.0.2, and I'm running out of patience trying all the possible permutations of rebuilding {clang,llvm,rust}.
<SiFuh_>
WTF? I can legally dump it in the air, water, or sewage systems but I need to inform them I am doing it if it less or equal to 1 radionuclide
<SiFuh_>
has gross alpha and gross beta concentrations in the leachate worked out under the TCLP that are each equal to or less than the concentration stated in the Australian drinking water guidelines for the radionuclide multiplied by 10.
<SiFuh_>
Mines way higher than that
<SiFuh_>
Mine's
<SiFuh_>
no person will receive greater than 300µSv a year as a result of the disposal
<SiFuh_>
So I guess I can dig a 6 foot hole in the back yard and bury it. Unless the river washes it out within 250 million years
<farkuhar>
Here's the reason the rust-bin make_footprint function needs further logic to prevent pkgmk MISSING mismatches. https://dpaste.com/8VS9SYWEU
<SiFuh_>
farkuhar: Do you guys over there say "In these, here woods". Instead of "In the woods, here"?
<farkuhar>
SiFuh_: "In the woods, here" is more common for me.
<SiFuh_>
Damn, I liked the other way
<farkuhar>
The other way isn't unheard of, it's just less common where I happen to live.
<SiFuh_>
Oh yeah, I kind of forgot you live on the Sphincter of DC
<SiFuh_>
Actually I was surprised to learn much about US I didn't even know from the Moonshiner's TV series
<farkuhar>
"mesa: 25.0.1 -> 25.0.2", now there's a port with the potential for breakage if you share the built package among machines with vastly different hardware. Just look at the README to get a sense of how versatile the configuration of that port is.
<SiFuh_>
For a country with so many droids walking around, the US has a surprisingly large amount of land that is still wild.
<SiFuh_>
Kind of puts that 'over-populated earth' bullshit into perspective.
<farkuhar>
Rebuilt rust, llvm, and clang, and I'm still seeing problems with gecko-profiler. process didn't exit successfully: `/usr/pkgmk/work/firefox/src/firefox-136.0.2/firefox-shared/release/build/gecko-profiler-434e4de64415f7fa/build-script-build`
<farkuhar>
Unable to find libclang: "the `libclang` shared library at /usr/lib/libclang.so.19.1.7 could not be opened: Dynamic loading not supported"
<farkuhar>
Sorting the firefox-shared/release/build/ directory by last-modified time, it appears that gecko-profiler is always the target that stops the build job. Now I should poke around the upstream site to see if that particular code got significant changes between 136.0.1 and 136.0.2
<farkuhar>
For good measure, I could also try the same build on CRUX-glibc, to see if it's a problem specific to musl.
<SiFuh_>
It doesn't matter what you talk about farkuhar never diverts from the path
<farkuhar>
Actually an easier solution might be `echo "ac_add_options --disable-gecko-profiler" >> .mozconfig`, but I don't know if that would have side effects like the slower startup that serpente noticed after the 135 -> 136 version bump.
<SiFuh_>
See what I mean?
<farkuhar>
I have to keep reminding myself that serpente observed slower startup with LibreWolf, not mainline firefox. But anyway, "ac_add_options --disable-gecko-profiler" didn't work, somehow it still got added to the list of build targets.
<SiFuh_>
He knows what you said before and the URL you posted and was already aware of it.
<farkuhar>
The unwanted activation of the gecko-profiler target seems to stem from tools/profiler/rust-api/build.rs, but I would have expected this codepath to be deactivated too, by "ac_add_options --disable-profiling" in mozconfig.
<farkuhar>
Anyway, in order to test whether the same problem arises on CRUX-glibc, I have to satisfy a few dependencies first (nodejs and cbindgen). If it's not a problem on CRUX-glibc, I'll just push the update and figure out a workaround for MUSL later.
<SiFuh_>
farkuhar: I wanted to bottle 100 beers today. But it seems the dog not eating for 5 days is enough to disrupt my schedule.
<SiFuh_>
So it looks like a vet trip to keep the woman happy
<SiFuh_>
It's not like she is starving. She does pick at the food and takes what she wants.