<braewoods>
Interesting. The turing smart screen doesn't load properly on about half of the unix clones I tested it on. I mean, the kernel driver complains about something.
<braewoods>
Haiku loads but complains about something being broken?
<braewoods>
Huh. Seems I found the problem. The turing smart screen only supports being written to?
<Begasus>
and that probably would be wrong too (doubt the irrxml version is the same as for assimp)
<OscarL>
someone didn't check if libIrrXML was already provided :-P
* Begasus
hides :)
<OscarL>
lol, from #5101: "Is there any benefit on using this one (guessing it's as a requirement for assimp?)?"
<OscarL>
at least you asked!
<Begasus>
probably dome work from GCI :)
<Begasus>
some*
<Begasus>
jikes, required for Qt5 (assimp)
<Begasus>
not bumping version there :P
<Begas_VM>
k, so far for the things installed on 64bit things are OK :)
<Begas_VM>
302 static ones (mostly boost, clang, llvm and qt)
<Begasus>
no patches upstreamed to assimp :/
<Begasus>
do we still suffer for the 2GB memory exhaustion on 32bit?
floof58 is now known as Guest12398
floof58 has joined #haiku
<OscarL>
I only got 2 GB of RAM on this VM, so... :-P
Guest12398 has quit [Ping timeout: 480 seconds]
<OscarL>
I think I only ran into an OOM for the first time (compiling something while Web+ was idling in the background) 5 days ago: "virtual memory exausted: Out of memory".
<Begas_VM>
seen that a few times, on the other laptop 6GB
<Begas_VM>
just for the assimp korli added a patch for that
<OscarL>
meanwhile... my "let's patch AboutSystem's replicant real quick" is getting bigger than expected :-/
<Begasus>
what's the patch doing?
<Begasus>
ps, the static lib can't be turned of by assimp
<Begasus>
doesn't take the system one for irrxml
<OscarL>
Attempting to fix a corner case... If you have an AboutSystem replicant on your desktop, and change CPUs (or move your Haiku drive to another PC)...
<OscarL>
it doesn't updates the CPU info.
<Begasus>
as in?
<OscarL>
This would be my 3rd time doing similar fix :-D... first it was the hrev not updating, then the kernel date, now it is the CPU name/frequency and number of cores.
<Begasus>
I don't see you doing that on bare metal so quickly :)
<OscarL>
To be fare, I was aware that that could happen, (even mentioned on my previous gerrit review)
<OscarL>
(s/fare/fair)
<OscarL>
Begasus can happen really easy if you're using an install on an USB drive.
<Begasus>
that too, switching devices
<OscarL>
that's the case I didn't considered earlier, and thus decided it wasn't worth fixing at the time :-D
diver1 has joined #haiku
Diver is now known as Guest12401
diver1 is now known as Diver
<Begasus>
like if I would hook up my drive on this laptop the CPU info wouldn't be correct either?
<OscarL>
by the way... thanks for mentioning the VM case.. now I can test that part :-D
<Begasus>
lol
<OscarL>
(didn't occured to me I could just change the number of cores in the VM to test :-D)
<Begasus>
feature request, remove border from the replicant :)
Guest12401 has quit [Ping timeout: 480 seconds]
<OscarL>
that will cost extra :-P
<OscarL>
and probably more of a task for Skipp_OSX.
<Begasus>
hmm ... that's what you get for helping out ... :P
<braewoods>
OscarL: that could be triggered on real hardware with something as simple as a CPU swap.
jmairboeck has joined #haiku
<braewoods>
Not very often but it does happen when people upgrade or similar.
<OscarL>
braewoods, yes, that's what happened on the image I`ve linked.
<OscarL>
well, no.... that seems more like a full mobo+cpu swap, I guess :-D
<braewoods>
What causes it? Caching data?
<OscarL>
(considering the cpu gens?, I'm not up to speed on intel gens)
<braewoods>
It sounds like it since if it was refreshed every boot it shouldn't happen.
<OscarL>
braewoods: just the way replicants work.... they are "archived" views (into flattened BMessages, I guess?)...
<Begas_VM>
new fix to the beta?
<braewoods>
Oh.
<Begas_VM>
hrev up to 79 now
<OscarL>
braewood, and in this case, that info was only updated on the regular app constructor, but not on the one for the replicant (where it gets "unarchived").
<OscarL>
braewoods: USB-serial displays, right? Good luck with it!
<braewoods>
of which maybe Haiku can run it well enough.
<braewoods>
OscarL: yea. i'm doing a low level abstraction and leaving higher level details for clients.
<braewoods>
like dealing with pixel formats and the like.
<jmairboeck>
Begas_VM: I requested that because it likely fixes the fontforge crash during the lilypond build. If it does, we will still need an actual buildmaster update, I guess ...
<Begasus>
thanks OscarL :)
<Begasus>
yeah jmairboeck
<braewoods>
OscarL: Not just usb serial. There's also some that need custom drivers. My project is for anything that uses USB to control an LCD.
<OscarL>
Begasus I swear there's a way to subscribe to those changes via .rss, but now I can't find the link, lol.
<braewoods>
Assuming the protocol can be discovered.
<Begasus>
so buildmaster is 10 revisions behind now
<braewoods>
Right now I'm working on the serial code for USB ACM devices. I'm finding some kernels are stricter than others. Notably one of my screens is rejected by FreeBSD and NetBSD kernels.
<OscarL>
braewoods: sounds awesome! I wish I could aford to play more with that type of hardware.
<braewoods>
OscarL: these hardware units are more affordable than you might think. ax206 or turing models are usually around 30 USD or less.
<OscarL>
(wouldn't help with my focus issues, but... would be cool)
<braewoods>
Or maybe you can pickup an old digital picture frame to hack into an ax206 dispaly.
<braewoods>
display*
<OscarL>
braewoods I might be poorer than expected :-) I live on 50/60 USD a month, hehe.
<braewoods>
Well, you're not missing a lot I guess. I realized that my end result project will mostly be Linux only but the low level display controls can be very portable.
<OscarL>
jmairboeck that's the one! Thanks.
<braewoods>
Mostly because of hardware sensor support being non-portable which is an input often used to these screens.
<braewoods>
So I decided to split my project in two so maybe someone else can use the portable code for something more advanced for their favorite unix clone.
<braewoods>
Plus the fact I'm using C to do all the processing and data moving will help reduce latency from the host end.
<jmairboeck>
OscarL: I just found out that cgit also provides one :)
<OscarL>
(and "h=master" for master)
<braewoods>
It's kinda weird seeing people using python to convert pixels to RGB565 format. Not a very effecient way to do that.
<braewoods>
It works but I can't help but think it slows down stuff.
<OscarL>
braewoods: that's the issue with languages that are good at being "glue"... you'll get people sniffing it too much :-P
<braewoods>
There's only one real advantage I can see to using that. Native code that isn't part of the standard distribution of Python is a pain to bundle with interpreted code.
<OscarL>
Python being an excellent glue, gets used in all kind of ways it really shouldn't, but... sometimes it makes things SO much easier..
<OscarL>
... that you can overlook the performance penalty.
<braewoods>
Well in my case, my client library may end up doing all the intense processing in itself so the interpreted code just has to pass data around.
<OscarL>
fantastic.
<braewoods>
In theory, we'll just see. Some features I may emulate in software. Some of these displays handle orientations natively so you don't need to manually rotate stuff in software.
<braewoods>
It's also funny how byte order of pixels varies between devices.
<Begas_VM>
errors during rebuild of current ldap
<Begas_VM>
will add conflicts when the new version builds ok :)
diver1 has joined #haiku
Diver is now known as Guest12407
diver1 is now known as Diver
Guest12407 has quit [Ping timeout: 480 seconds]
mmu_man has joined #haiku
<Begas_VM>
run configure with --enable-mdb to run MDB tests ... I did (stupid openldap)
HaikuUser has joined #haiku
<OscarL>
heh
<trungnt2910[m]>
For hpkg files, what does package:install-path do?
<trungnt2910[m]>
It seems to be undocumented.
jmairboeck has quit [Ping timeout: 480 seconds]
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
<Begasus>
they removed the _r ones in (atleast in the 2.6.* versions I've checked)
<OscarL>
man1 looks similar to mani (Spanish for peanuts), now I'm hungry.
Diver has quit [Read error: Connection reset by peer]
<Begasus>
heh
Diver has joined #haiku
<OscarL>
Mmm, that mail I linked suggest that using both version of the libldap libs can cause problems (due to the lack of symbol versioning)... wonder if that's still an issue.
<OscarL>
seems that mail asking for advice got not answers, lol.
<Begasus>
yeah, no reply there :)
mmu_man has quit [Ping timeout: 480 seconds]
Diver has quit [Quit: Leaving.]
* Begasus
bangs head ...
Diver has joined #haiku
Diver is now known as Diver
mmu_man has joined #haiku
<Begasus>
had the wrong recipe open for unzip :)
<OscarL>
happens to the best of us. (not to me, of course, no sir!)
<OscarL>
Something's broken with that. I've never used it before, but after clicking on "View-Show Bookmark bar"... it hide the emtpy Bookmark Bar I had, and now it won't show it back.
<OscarL>
(even after I copied some bookmars on the "Bookmark Bar" folder)
<botifico-c849d97b>
[haikuports/haikuports] Begasus f8db1d5 - bzip2, remove empty man dir (#8548)
<Begasus>
bbl
Guest12423 has quit [Ping timeout: 480 seconds]
<OscarL>
So... no inverse of /bin/message that I can find, but... we do seem to have a "BJson" class that can do Json->BMessage. Just no round-trips with either, it seems.
<OscarL>
Adding to the infinite ToDo/Wish list: turn BMessage::PrintToStream() around into BMessage::ReadFromStream()
<OscarL>
Now that I read BrunoSpr's screenshot properly... he has "Bookmark bar" *outside* of the "Bookmarks" folder... that might explain why it removing the "Application" file doesn't help for him.
mbrumbelow has quit [Quit: Vision[]: i've been blurred!]
B2IA has quit [Quit: Vision[]: i've been blurred!]
<botifico-c849d97b>
[haikuports/haikuports] korli d1a4c18 - mgba: bump version
OscarL has joined #haiku
mbrumbelow has joined #haiku
talos has quit [Quit: WeeChat 3.8]
talos has joined #haiku
xet7 has joined #haiku
<PulkoMandy>
OscarL: the PrintToStream format is not always complete enough to allow reconstructing the complete message
<PulkoMandy>
And I think the json one can work both ways, but only for a subset of BMessage that can be stored as json (we couldn't make it be fully generic)
<OscarL>
re PrintToStream... understandable. I was just wishing for a quick way to "fix" a flattened BMessage :-D
<OscarL>
(in this particular case to enable the "show bookmarks bar" on Web+ Application settings file)
tuaris has joined #haiku
<augiedoggie>
i think xmlbmessage is lossless
<augiedoggie>
i also have a jsonbmessage cli tool hiding in my projects folder somewhere
<OscarL>
augiedoggie, thanks for the xmlbmessage tip. lets's "pkginstall cmd:xmlbmessage" to try it :-)
Begas_VM has quit [Quit: Vision[]: i've been blurred!]
<Begasus_32>
re
<OscarL>
wb
Begas_VM has joined #haiku
<OscarL>
augiedoggie, welp... xmlbmessage just crashed on me :-D
<OscarL>
that's on stdout. crash report shows an __assert_fail.
gouchi has joined #haiku
<mbrumbelow>
waddlesplash: You can wack the old Paladin repo. I took a look and it looks good.
gouchi has quit [Remote host closed the connection]
<Begasus_32>
man libsolv
<Begasus_32>
man: No entry for libsolv in the manual.
<OscarL>
Begasus: that extra man... very manly! :-P
<Begas_VM>
yeah, no wonder you cant consult the manpage :P
<OscarL>
augiedoggie: if you uncover that jsonbmessage tool... I would certainly like it more than I like xmlbmessage (or xml in general :-D). No hurries, thou.
<OscarL>
ok... time to reboot to test if my fix for AboutSystem actually works when moving to different PCs...
<waddlesplash>
well, guess I get to try and bisect this one today
<Habbie>
ouch
jmairboeck has quit [Quit: Konversation terminated!]
<OscarL>
re: lpeg.so... its a similar case with Python module extensions when they are compiled into .so files. We fought a similar issue on the "unicorn" recipe with Begasus.