erysdren has quit [Quit: Konversation terminated!]
HaikuUser3 has quit [Quit: Vision[]: i've been blurred!]
mmu_man has quit [Ping timeout: 480 seconds]
OscarL has joined #haiku
<OscarL>
yikes! Updated my 32 bits install to beta5... first git started to "speak" Spanish, even if locale language is set to "en", and now it crashed :-(
<OscarL>
segment violation on a "memcmp", near "bsearch_hash"
rakka has quit [Remote host closed the connection]
rakka has joined #haiku
nipos has left #haiku [Disconnected: Replaced by new connection]
nipos has joined #haiku
<jpelczar>
waddlesplash: I am changing this for my own fun or experiments. Not everything is necessarily production ready.
<waddlesplash>
jpelczar: ok
<waddlesplash>
OscarL: I made a fix to 32bit that only made it in to the release branch today
<Begasus>
openmpi is disabled also it seems OscarL
<OscarL>
right
<Begasus>
crossing fingers PulkoMandy :)
<Begasus>
9.6M, not that big ... launching build
<OscarL>
welp... only transmission remains to be checked before dropping libevent21
HaikuUser3 has joined #haiku
<OscarL>
Begasus: I think we should just drop the "www-aps/seeks" recipe. For a "social websearch" seems pretty dead (last source activity in 2012). Recipe doesn't builds (I tried last year). https://repology.org/project/seeks/versions
<Begasus>
let's see how openmpi goes, if so I can do a check on Transmission too
<Begasus>
+1 then :)
<Begasus>
put a notice on the forum for it? (comes to mind I opened a thread for removals some time ago)
<OscarL>
recipe has been disabled for more than a year already,
<OscarL>
and before that... I couldn't even find a .hpkg on depot.
<Begasus>
ah, then it shouldn't be an issue
<OscarL>
so unless someone has a build from the .bep era...
<Begasus>
someone wanting to tackle it can still find it in the history
<OscarL>
indeed.
<OscarL>
or start afresh... after they first revive upstream :-P
<Begasus>
maybe I'll just grab residualvm and put it up on a github repo (for backups) and nuke it at haikuports
<Begasus>
k, biab
freddietilley has joined #haiku
Jixt has joined #haiku
<Begasus>
openmpi and transmission build fine (64bit)
<Begasus>
g'morning Jixt
<Jixt>
good morning Begasus.
<OscarL>
cool! then you can drop libevent21 then (and tick its mark on #10865) :-)
* OscarL
is trying to see if he can build the last recipe using old libevent 1.x
<Begasus>
yeah, first drop old one, then I don't have to set a version string on these
<OscarL>
opened PR for dropping two recipes using old libevent 1.x... this last one I don't do the same because it is a native app :-D
<OscarL>
mmm Torrentor crash was on "generating keys", surely doesn't has *anything* to do with my awesome changes in crypto.c file!
<OscarL>
Begasus: because I wasn't even aware I didn't had that fix in yet.
<OscarL>
(till waddlesplash mentioned it earlier)
<OscarL>
and I get distracted way too easy... one thing at the time.
<Begasus>
on 64bit I don't experience weird crashes (me knocks on wood)
<Begasus>
heh
<OscarL>
also... hard to build stuff if git and bash keep crashing.
<Begasus>
make sure your branch is uptodate, then changes are not that big?
<OscarL>
yea, that issue affects mostly 32 bits it seems.
<Begasus>
ah, right, got some there
<OscarL>
Begasus: I would need to build all of haiku.hpkg in any case, so.
<Begasus>
haiku+haiku_devel should be enough I think?
* OscarL
goes back to Torrentor! and disables "Use Encryption" :-D
<OscarL>
I have never build haiku for 32 bits, would need to setup things... I rather wait one day more.
<Begasus>
launching 32bit build laptop ...
<Begasus>
Would be nice if softwareupdater had an option to disable package update (individual package)
Anarchos has joined #haiku
<Begasus>
still at: Building [=======================> ] 356/361: cargo
<Begasus>
Hi Anarchos
<Anarchos>
hello Begasus
<Anarchos>
Begasus i just wake up one hour ago
<OscarL>
holy cow, it is downloading beta4.iso torrent!
<Begasus>
sleeping in Anarchos? :)
<Begasus>
congrats OscarL! :D
<OscarL>
well, it started to do so at least... now seems hang (before it, already had a core pegged to 100%)
<Begasus>
rebooting 32bit after update
<Anarchos>
Begasus i Had much sleep missing last days.
<Begasus>
guess nexcloud and falkon will be broken (backed up previous qtwebengine)
<OscarL>
Tracker crashed badly.
<Begasus>
back to base then :P
<OscarL>
me forces a reboot.
<Anarchos>
OscarL i experience some freeze of the keyboard/mouse recently too
* Anarchos
thinks he will start to do small recurring donation to haiku inc this year
<OscarL>
Begasus: to be fair... this is hrev is known to have issues, the iso was larger than the available space, and then we have my hacks on top :-D
<OscarL>
Anarchos: you use the 32 bits version on that netbook?
<OscarL>
(good day to you, btw :-D)
<Anarchos>
OscarL 64 bits version on the other laptop
<OscarL>
I see.
<OscarL>
on the Atom N450 netbook I used to have to disable SMP *all* the time. Now it works most of the times, but sometimes has crashes that go away with SMP disabled.
<Anarchos>
OscarL i never had to disable smp since a very long time !
<Anarchos>
OscarL woks perfectly here
<OscarL>
glad it works for you! it must be some poor BIOS on the netbooks I have that causes problems then.
HaikuUser3 has joined #haiku
<Anarchos>
OscarL tried a bios update ?
<OscarL>
impossible to find any BIOS file for these ones.
<Anarchos>
OscarL ok...
<OscarL>
well, not *impossible*, but almost.
<OscarL>
and I rather do not risk bricking them :-)
<Anarchos>
OscarL sure
<Anarchos>
OscarL bios ninjutsu is a fine and dangerous art :)
<OscarL>
I remember "hot-swapping" BIOS in the old days... that was scary :-D
<Anarchos>
OscarL why should it be ? When the computer is running you can get rid of the bios, no ?
<OscarL>
from the software side of things... no problem...
<OscarL>
electrically... you can fry the motherboard, or parts of it.
<Anarchos>
extract it with plastic tools ?
<OscarL>
that will help avoid short-circuits, still can break stuff for breaking, say, ground connection from one side, while other parts are still recieving power.
<OscarL>
I broke something on one of my motherboards that way. It lost the ability to retain RTC/BIOS settings in case of power-cuts.
<Anarchos>
ah ok
<Anarchos>
if i try, i will keep in mind to wire the ground pin and disconnect it last :)
pabs has quit [Quit: Don't rest until all the world is paved in moss and greenery.]
<phschafft>
good morning.
<dovsienko>
hi
pabs has joined #haiku
<OscarL>
o/
Begasus has quit [Ping timeout: 480 seconds]
<OscarL>
Begasus: just noticed that Torrentor! appears to have been written by an Argentine (about 500 Km away from where I am, thou... and more than a decade ago :-D)
Jixt has quit [Ping timeout: 480 seconds]
HaikuUser3 has quit [Quit: Vision[]: i've been blurred!]
Begasus has joined #haiku
jpelczar has quit [Quit: Konversation terminated!]
<Begasus>
back ... for how long? :P
mmu_man is now known as Guest2355
mmu_man has joined #haiku
<dovsienko>
OscarL: I've been to Argentine once, but it was on a busy business so I only had a few glimpses of Buenos Aires city centre, let alone the rest of the country
<Anarchos>
i wonder if a PR to add 'alias reboot = shutdown -r' may be accepted.
<OscarL>
dovsienko: :-). I try to stay away from Buenos Aires as much as I can (not a fan of such large cities). Hope you had at least a decent time there.
Guest2355 has quit [Ping timeout: 480 seconds]
<OscarL>
only time I was there for more than just passing by in a bus en-route to somewhere else, was work-related too (full week).
<OscarL>
Anarchos: it might. KDL even has reboot as a command.
<OscarL>
if you get a -2... point to that command and say... it is for consistency's sake!
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<OscarL>
guess he did tried that alias.
Anarchos has joined #haiku
<Anarchos>
sorry i experience one ot this keyboard freeze: killing input_server didn't help so i had to reboot
<Begasus>
no good with that alias also then :)
<Anarchos>
issues not related :)
<Begasus>
;)
<Begasus>
Building [=======================> ] 356/361: cargo
<Begasus>
no progress still :(
<Anarchos>
Begasus what are you building ?
<Begasus>
cargo-c
<Anarchos>
ah ok
<Begasus>
rust is always a pain to tackle
<Anarchos>
Begasus i find all those annotations in rust very messy : pub fn mut & ..... it makes source code really unreadible
<OscarL>
"At least it ain't Perl!" (TM)
<Coldfirex>
we have working rust?
<OscarL>
even on 32 bits... "pkgman search rust" -> rust_bin_x86
<Begasus>
it's working fine (for my needs at least)
<Begasus>
mostly* :D
MisthaLu has quit [Quit: Leaving]
xet7 has joined #haiku
<Coldfirex>
Nice. Ive never used it just didnt realize
<Begasus>
nuked it, takes me too long :P
_-Caleb-_ has left #haiku [#haiku]
wicknix has joined #haiku
wicknix_ has quit [Ping timeout: 480 seconds]
_-Caleb-_ has joined #haiku
<OscarL>
Begasus: good news! Seems my silly libtransmission/crypto.c changes won't be needed after all. Found that the original author of Torrentor has one commit (not present on HaikuArchives/Torrentor) updating libtransmission to 2.93, and that seems to solve the incompatibilities with openssl3 :-)
<OscarL>
need to see why it crashes on exit, though.
<OscarL>
seems libevent related.
<OscarL>
Task for later, now... sleep. See you around, folks. Have a good one!
OscarL has quit [Quit: zzzZZZzzz]
HaikuUser3 has joined #haiku
akashkumar has quit [Quit: Connection closed for inactivity]
diver has quit [Read error: Connection reset by peer]
diver1 has quit [Read error: Connection reset by peer]
diver has joined #haiku
HaikuUser3 has quit [Quit: Vision[]: i've been blurred!]
coolcoder613_beos has quit [Ping timeout: 480 seconds]
tuaris has joined #haiku
<hsp>
my recycle bin has been crashing recently when I empty it. anyone have any ideas?
<gordonjcp>
hsp: weirdly so has mine, but my physical recycling bin
<gordonjcp>
it turns out the little circlip thing holding the wheel on had fallen off
<gordonjcp>
hsp: are you short of disk space? you might be in the paradoxical situation of having not enough space to delete stuff
<hsp>
50G free, that should be enough :)
<Anarchos>
hsp did you run checkfs ?
<hsp>
no
bbjimmy has quit [Quit: Vision[]: i've been blurred!]
bbjimmy has joined #haiku
<hsp>
checkfs no errors
<Anarchos>
hsp there is some stall when dealing with many small files, but i dunno if it explains a crash
mittwerk has joined #haiku
<hsp>
it would be better to have a menu item for direct deletion instead of moving to the recycle bin
<PulkoMandy>
there is one but you have to press shift while opening the menu I think
<PulkoMandy>
(so you don't delete files by accident)
_-Caleb-_ has left #haiku [#haiku]
<hsp>
nothing shift
_-Caleb-_ has joined #haiku
<hsp>
this does not work with shift either
<augiedoggie>
yeah, i don't see any code that would do that in Tracker
<augiedoggie>
the only Delete option is for things already in the trash
<Anarchos>
hsp SHIFT + Delete
<Anarchos>
augiedoggie shift+delete in tracker does it.
<augiedoggie>
but the question was about having a menu item to do it
<augiedoggie>
holding shift doesn't change the menu text
<hsp>
but should be feasible
<Anarchos>
hsp but it would be redundant ?
HaikuUser3 has joined #haiku
<Anarchos>
how uses leak_analyser.sh ?
Nephele has joined #haiku
Nephele is now known as nephelXmpp
nephelXmpp is now known as nephele_xmpp
nephele_xmpp has left #haiku [#haiku]
nephele_xmpp has joined #haiku
<nephele_xmpp>
Does anyone have experience with Haiku on 240hz or 480hz displays?
<nephele_xmpp>
As in, anyone know if those would be expected to work?
<PulkoMandy>
it depends on the video driver you want to use
dby has joined #haiku
<PulkoMandy>
on Intel cards (the driver I know best) I think it would need a higher bandwidth and so switching to dual channel displayport or whatever the equivalent is in HDMI
<PulkoMandy>
I don't think we have this working well at the moment (even using non-60Hz on VGA does not seem to work reliably on current hardware :( )
<nephele_xmpp>
Hmm, my current desktop has a ivy bridge card (according to Screen) and two displayport ports. I guess I have to check what those even support?
<nephele_xmpp>
The computer spec sheet does not say what these ports are capable of (as in what version they have) darn
illwieckz has quit [Ping timeout: 480 seconds]
HaikuUser3 has quit [Quit: Vision[]: i've been blurred!]
illwieckz has joined #haiku
HaikuUser3 has joined #haiku
bjorkint0sh has quit [Quit: "Every day, computers are making people easier to use." David Temkin]
<nekobot>
[haiku/haiku] 95a871b37d00 - libroot: Handle invalid timeouts being passed to unnamed_sem_timedwait.
<nekobot>
[haiku/haiku] 797b4eef8bfe - pthread: pthread_cancel() on non-existent thread should return ESRCH.
<nekobot>
[haiku/haiku] 48126360286a - pthread: Return EINVAL for pthread_detach() if the thread is already detached.
<nekobot>
[haiku/haiku] 2301efc9d944 - pthread: Make rwlocks keep waiting on EINTR.
<dovsienko>
it is difficult to remember, but I guess BeOS didn't have it
<dovsienko>
that said, in some future it would be nice to have both multiple users support and a login screen
<theotheroracle>
hm alright, is that planned to be implemented before stable ? it seems like an important feature to have for launch
mmlr has joined #haiku
<dovsienko>
I do not prioritise new features in this OS, but you can raise the idea
<theotheroracle>
fair enough
<dovsienko>
from a software developer point of view, in some projects there's more value in fixing bugs than in implementing new features
<theotheroracle>
i understand
<Anarchos>
theotheroracle beos and haiku are meant to be single users.
<Anarchos>
theotheroracle that's why there is no need for a login screen
<theotheroracle>
i'm more concerned about the security of it rather than having multiple users
<waddlesplash>
we eventually should have a login screen
<waddlesplash>
but it's not a priority at the moment
<Anarchos>
i remember a login window somewhere (and not screen saver) but can't remember more
<dovsienko>
workaround: start the screensaver at boot :-D
<Anarchos>
waddlesplash i could not get any trace from the acpi_als driver. do you know who could explain me how to force acpi drivers to init ?
<waddlesplash>
it depends on what kind of driver it is and at what stage it gets loaded
<Anarchos>
waddlesplash i activated TRACE in acpi_als.cpp
dby has quit [Quit: Vision[]: i've been blurred!]
Anarchos has quit [Quit: Vision[]: i've been blurred!]
Anarchos has joined #haiku
<Anarchos>
were they some modif in PS2 driver ? My keyboard stops working after one hour.
<waddlesplash>
hello mmlr
OscarL has joined #haiku
<OscarL>
oh! me likey hrev58061 and hrev58062! I hope those will at least improve some failing tests on the Python's datetime tests (or at least allow me to narrow down the issue).
<OscarL>
mmm, "cgit" is missing the hrev58063 tag :-/
theotheroracle has quit [Quit: Vision[]: i've been blurred!]
<Anarchos>
no latex for me tonight : texmf.cnf not found...
SLema has quit [Quit: Vision[]: i've been blurred!]
SLema has joined #haiku
<Anarchos>
waddlesplash isn't it too difficult to switch between architectures when you develop ?
<waddlesplash>
Anarchos: those aren't my changes
<waddlesplash>
I just clicked the "merge" button
<Anarchos>
what a great work :)
<waddlesplash>
mmlr: so, unrelated to all the recent stuff: one of the main ways we get into trouble in low memory situations is that we don't reserve memory for page_mapping objects
<waddlesplash>
so we can wind up in a page fault and have a valid page reservation with a page available but the page_mapping object_cache is out of objects and can't get more
<waddlesplash>
one of the key reasons this happens is because the low_resource system flushes object_caches when memory gets tight
<waddlesplash>
IMO we should make the CACHE_PRIORITY_VIP flag have the low_resource system not flush caches, which would help here
<mmlr>
can't we add a concept of "minimum reserve" to the object_caches?
<waddlesplash>
we could do that too, yes
<mmlr>
and then not only reserve pages but also reserve mappings there?
<waddlesplash>
but I also think we should try to use page_mapping objects less. In particular I don't think we need them for the standard case of a single area mapped into a single process space
<waddlesplash>
we don't need them for LOCKED memory, we should not need them for pages that are fully "owned" by one area
<waddlesplash>
if and when the area gets cloned or shared, then we will have to allocate a lot of page_mapping objects then and "convert" it to use such shared mappings
<mmlr>
where do we need them really?
<waddlesplash>
we need them when iterating over all the places a page is mapped
<waddlesplash>
but if a page is mapped only once, with one cache and one area, then this is easy enough to get
<waddlesplash>
it
<waddlesplash>
's when there's a whole tree of caches and areas that traversing them to see who's got what mapped becomes impossible
<mmlr>
but they can reside in multiple layers of caches as well, even when just in a single owner
<waddlesplash>
in that case we would also have to use a mapping object, sure
<waddlesplash>
basically I am just suggesting we eliminate the use of the objects when there is 1 cache and 1 area
<waddlesplash>
anything that will create a second cache would have to switch to objects
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<waddlesplash>
this would resolve page faults from heap accesses leading to OOM causing problems at least
<waddlesplash>
but yes, "minimum reserve" in object caches sounds necessary to solve this really
<mmlr>
how would you keep track of the 1:1:1 mapping case?
<waddlesplash>
mmlr: looks like we already have object_cache_set_minimum_reserve. however it doesn't check for failure and it just sets
<waddlesplash>
probably a bit in the page
<waddlesplash>
or maybe reuse fWiredCount or something
<mmlr>
sounds fragile
<waddlesplash>
anyway, so I guess we want to add object_cache_add_minimum_reserve() and then add/subtract in the right places
<mmlr>
finding all the places where the conversion needs to happen
<waddlesplash>
the conversion needs to happen with whatever accesses "mappings" object in vm_page
<mmlr>
when it notices that there aren't any mapping objects?
<waddlesplash>
but yes, that's not a small number of places
<waddlesplash>
mmlr: we already don't use mapping objects when we have locked memory
<mmlr>
interesting
<waddlesplash>
pretty much all kernel areas are locked and thus no mappings are used
<mmlr>
does any realistic Haiku setup need 128PB of RAM?
<waddlesplash>
probably not, we actually disable everything above 512GB at the moment
<waddlesplash>
and in fact there are tickets about systems with 128GB failing to boot with panic can't allocate early page
<waddlesplash>
is there something else in LA57 that's useful perhaps? not sure why korli implemented it otherwise
<phschafft>
thanks for the link.
Coldfirex has quit [Remote host closed the connection]
coolcoder613_beos has joined #haiku
mmlr has quit [Quit: Vision[]: i've been blurred!]
nephele_xmpp has joined #haiku
<Oscar-L>
waddlesplash: I've beein getting "Call (BTimeSource::PerformanceTimeFor: performance time too large)" crashes for media_addon_server, both on VBox, and on bare metal. Couldn't find a ticket for this yet. Should I open one?
<Oscar-L>
(beta5, 64 bits)
<waddlesplash>
yeah, possibly
<waddlesplash>
it's interesting this happens on VBox
<waddlesplash>
this should only happen when the audio driver stalls for a long time
<Oscar-L>
my hardware being too slow might have something to do with that? :-D
<waddlesplash>
unlikely
<waddlesplash>
it would have to get no interrupts for at least 10 seconds
<waddlesplash>
I think a lot more actually
<Oscar-L>
ok. Will collect the .reports I have from the different machines and open a ticket then.
nephele_xmpp has left #haiku [Error from remote client]
nephele_xmpp has joined #haiku
nephele_xmpp has left #haiku [Disconnected: Hibernating too long]