<coolcoder613>
So, Fiwix can't boot from the CDROM on my machine for whatever reason
<coolcoder613>
And I wanted to install it
<OscarL>
phschafft: presence of "export SETUPTOOLS_SCM_PRETEND_VERSION=" in HaikuPorts tree shows that there are at least 16 python package authors that suck at this :-)
<coolcoder613>
So I (or Mikaku) had the idea to use a connect the disk to a different maachine
pbsds has joined #haiku
<coolcoder613>
I took it out, and... it's 2.5" IDE which I do not have an adapter for
<coolcoder613>
So, I thought a bit
<coolcoder613>
And hit upon the idea of using DD
<phschafft>
OscarL: we had some discussions at different projects on how to include both data in some nice way. but just pretenting there is no release only git is clearly the wrong way.
<coolcoder613>
First, I would use DD to backup the HD onto a USB
<OscarL>
phschafft: combine that with haikuporter creating a .git for you...
<coolcoder613>
Then, I first thought to just create an image the same size as the HD, but to play it safe, I installed on to a copy of the HD backup image
<OscarL>
lots of packages with "0.1~dev" (unless that SETUPTOOLS_SCM_PRETEND_VERSION override is used, or haikuporter instructed to NOT create the git repo)
<coolcoder613>
Now, I have copied this installed image to an ext2 USB
<coolcoder613>
and i am going to use a Linux LiveCD to DD the image on the USB to the HD
<phschafft>
OscarL: ... I see...
* phschafft
listens to coolcoder613 in the mean time.
B2IA has quit [Quit: Vision[]: i've been blurred!]
<phschafft>
don't go there unless you have nothing to do for the next few weeks ;)
<scantysnax>
haha, just checked it out, bookmarked.
<phschafft>
one more lost soul.
<scantysnax>
lost my soul a long time ago.
mmu_man has quit [Ping timeout: 480 seconds]
<coolcoder613>
DD done, booted Fiwix and IT WORKS!!!
<scantysnax>
Fiwix?
mmu_man has joined #haiku
* coolcoder613
tells scantysnax to look at the IRC logs
<OscarL>
I feel cheated... you said it would take all day and night! :-P
* OscarL
checks clock and calendar, just in case he lost a day again.
<coolcoder613>
Who knows, maybe the bs=10m vs bs=4m helped, or busybox dd on linux vs gnu dd on BeOS, or a good part of the image being the same as the disk...
dby has quit [Read error: Connection reset by peer]
* coolcoder613
unscrewed a a keyboard, saw rubber domes missing
<coolcoder613>
s/a a /a
Diver has quit [Quit: Leaving.]
Diver has joined #haiku
erysdren has quit [Quit: Konversation terminated!]
<Begasus>
wb PulkoMandy, no major stuff missed (and there are still the logs) :)
<Begasus>
Warning: POLICY WARNING: "/packaging/nodejs20/bin/node" needs library "libbrotlidec.so.1", but the package doesn't seem to declare that as a requirement
<Begasus>
k, looks like we can skip that build part for nodejs by including system brotli
* coolcoder613
waves to zard, tells zard about his new blog post
* zard
checks it out
* coolcoder613
congratulates zard an any and all progress on webkit
<zard>
Thanks :)
<zard>
Also, nice blog post
<Begasus>
OscarL, when you are up and had your coffee, v8 still using cmd:python :)
<Begasus>
Hi zard, wb coolcoder613 :)
<zard>
o/
<coolcoder613>
zard: Thanks
<Begasus>
resolving build dependencies failed: build-prerequires "cmd:python" of package "v8_x86-7.4.288.28" could not be resolved requires "lib:libv8_x86" of package "libresprite_x86-1.1~git" could not be resolved
<coolcoder613>
I hope to get better at writing though practice
<Begasus>
when did this become a show stopper?
<Begasus>
v8 is still available in the depot
DKnoto has joined #haiku
dodo75 has quit [Remote host closed the connection]
_-Caleb-_ has left #haiku [#haiku]
_-Caleb-_ has joined #haiku
tqh has joined #haiku
<Begasus>
ERROR at //build/config/BUILDCONFIG.gn:412:1: Unknown function.
<zard>
I got BeOS installed in a VM for the first time
<zard>
I want to compare BeOS
<zard>
's handling of B_BITMAP_ACCEPTS_VIEWS to Haiku's handling
<Begasus>
nice zard :) think I still have some vm for BeOS around ... somewhere* :)
<dalme_>
I've never tried BeOS either zard
dpirate has quit [Write error: connection closed]
dpirate has joined #haiku
OscarL has joined #haiku
<OscarL>
Can't do BeOS without OpenTracker anymore here.
<zard>
grr, my mouse disappeared. Let's see how long it takes till I have to restart
<zard>
What is OpenTracker?
<OscarL>
open-sourced version of Tracker, the one we use on Haiku (well, minus improvementes over the years).
<zard>
I wonder, can you replace Be's kernel with Haiku's, swap out the servers one by one with Haiku's, and so on until you have a system that is entirely Haiku?
<OscarL>
at the start, "OpenBeOS" was basically a collection of replacement apps/server/drivers.
<OscarL>
I ran quite the "FrankenBeOS" back in the day... R5 + every OpenBeOS replacement that got more or less ready.
<OscarL>
(before I switched to 5.1, at least)
<zard>
Hence why I wondered. I guess most of it comes down to whether the kernel supports running BeOS userspace on top of it or not
<zard>
Mmm, imagine a
<zard>
Mmm, imagine a "FrankenWindows" where half the things are replaced with ReactOS components :P
walkingdisaster has quit [Quit: Vision[]: i've been blurred!]
Oscar-L has joined #haiku
<zard>
Yes! BeOS's implementation of B_BITMAP_ACCEPTS_VIEWS doesn't clear the bitmap!
<zard>
Now I have a good argument for changing Haiku's implementation to do that as well
<waddlesplash>
zard: you used to be able to slowly replace things this way to get from BeOS to Haiku, but you can't anymore, not for a long time
* zard
has been trying to figure out a way to get B_BITMAP_ACCEPTS_VIEWS to preserve the existing contents of an area of memory
<PulkoMandy>
zard: Ideally, the bitmap on uiprocess side should not need to accept views. Thiswould save quite a lot of resources (starting a new background window thread for each bitmap)
xet7 has quit [Ping timeout: 480 seconds]
<PulkoMandy>
Also I'm not sure what you're comparing with, since in beos yo couldn't create a bitmap from an existing area? So there it's always starting from empty/uninitialized data?
mmu_man has joined #haiku
<zard>
I'm comparing it with a regular bitmap created with B_BITMAP_ACCEPTS_VIEWS
<zard>
Haiku specifically clears it to white as if it had the B_CLEAR_TO_WHITE flag
<zard>
I think that avoiding using B_BITMAP_ACCEPTS_VIEWS would require changing the API of WebKit
<zard>
Unless you just simple want to check if we are in UIProcess and avoid using the B_BITMAP_ACCEPTS_VIEWS flag in that case
<waddlesplash>
yes
<waddlesplash>
but also I think we should redo how we manage bitmaps in WebKit anyway, because we run out of system threads very often
<waddlesplash>
when on graphics-intense websites like Google Maps
<waddlesplash>
we should probably instead create the accept-views bitmaps on-demand (by recreating the bitmap if necessary)
<waddlesplash>
this would then work in UIProcess also
<zard>
mmm, that gives me an idea
<waddlesplash>
for that matter, we can create the BBitmap on demand
<waddlesplash>
although maybe we already do that
<zard>
Yes, we do
<waddlesplash>
hm
dalme_ has joined #haiku
<waddlesplash>
how does WebKit handle painting? does it have something like a Painter object?
<zard>
Ok, looks like it would be natural to only create a BBitmap with B_BITMAP_ACCEPTS_VIEWS when a graphics context is requested
<waddlesplash>
and then we ask to attach a Painter object to a view, right?
<waddlesplash>
right
<waddlesplash>
do we know that at bitmap creation time?
<waddlesplash>
maybe we didn't before but now we do? idk
* zard
tries to figure the big-picture overview how webkit draws
<zard>
So yes, we essentially ask it to draw the web page into a BView
<zard>
DrawingAreaCoordinatedGraphics::display has the code for that
_-Caleb-_ has left #haiku [#haiku]
_-Caleb-_ has joined #haiku
juanjo has joined #haiku
<Begasus>
re
<PulkoMandy>
All the drawing will be through graphicscontext, but it may not be immediately clear which bitmaps will be drawn on. So the current code always sets the "accepts views" flag, just in case. It can probably be safely removed in a lot of places
<Begasus>
Hi OscarL
<OscarL>
Hey there Begasus!
<OscarL>
Just finished building python_pkgconfig + borgbackup the way I think they should. Not sure I've made my intentions clear enough on the comments, so... I better show what I mean :-D
<Begasus>
think I'm done testing nodejs, no go on 32bit
<PulkoMandy>
But yes, I was thinking a "are we in ths ui process" hack for now. Just to get something displayed. I tend to try to make things work first, and then try to optimize when it's working and I can test and see what I'm doing
jmairboeck has joined #haiku
<Begasus>
Hi jmairboeck
<Begasus>
you can push to PR and see where it ends OscarL
<zard>
Here's what I'm thinking of: ShareableBitmap::createGraphicsContext creates a BBitmap anyway, so we may as well add B_BITMAP_ACCEPTS_VIEWS there. And ShareableBitmap::createPlatformImage will not create a BBitmap that accepts views.
<zard>
So just a slight tweak of what I was going to do anyway
<OscarL>
Begasus: they seem to have a missmatch on types for nodejs on that 20.x branch, yeah. for 21.x those files seem to have changed quite a bit. Not sure if it will help or not for 32 bits thou :-D
<Begasus>
np OscarL, the previous 20 version didn't work out on 32bit too
<Begasus>
tried nodejs16, but that failed also on 32bit
<PulkoMandy>
zard: Ok then, sounds good :)
<Begasus>
korli already replied to an older issue it's not worth the effort, would have been nice though
<Begasus>
just can't build qtwebengine with newer ICU as it needs nodejs
Begasus_renga has left #haiku [#haiku]
mmu_man has quit [Ping timeout: 480 seconds]
<Begasus>
catch up in the morning
<Begasus>
closing down here, cu peeps!
Begasus has quit [Quit: Vision[]: i've been blurred!]
MisthaLu has quit [Quit: Leaving]
zard has quit [Quit: leaving]
dodo75 has quit [Quit: Vision[]: i've been blurred!]
yann64 has joined #haiku
HaikuUser has joined #haiku
<yann64>
/help
<OscarL>
How may I assist you today yann64? Beep boop.
<yann64>
:D
<yann64>
Getting confused as to how I can identify my nick on the XMPP bridge
<OscarL>
no idea there, sorry. <weird clicking noices>
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
zardshard has left #haiku [Error from remote client]
<HaikuUser>
but I've executed export JAVA_INCLUDE_PATH2=/boot/system/bin
<HaikuUser>
for reference, trying to do a cmake on MARIADB :)
<OscarL>
that doesn't sounds right to me. /bin is already in $PATH, and I would think that JAVA_INCLUDE_PATH might be something like "/system/lib/openjkd17/" or something like that.
<OscarL>
first off... I'd make sure I got installed whatever version of java MariaDB requires (assuming we have them in Haiku).
<augiedoggie>
that's a cmake variable, not an environment variable
<HaikuUser>
I edit the file the MakeFile, and changed name to JAVA_INCLUDE_PATH and it worked !
<HaikuUser>
it is now bulding MARIADB :)
<HaikuUser>
thanks
HaikuUser2 has joined #haiku
vdamewood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
HaikuUser2 has quit [Quit: Vision[]: i've been blurred!]
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
<OscarL>
leason learned... do not run "git reflog expire --expire-unreachable now --all" if git stash is not empty :-/
mmu_man has quit [Ping timeout: 480 seconds]
tuaris has quit [Quit: Leaving.]
_-Caleb-_ has left #haiku [#haiku]
_-Caleb-_ has joined #haiku
mmu_man has joined #haiku
<OscarL>
that last "--all" *really* means all. Oh well, at least I found why I have lost some stashes in the past :-D
_-Caleb-_ has left #haiku [#haiku]
_-Caleb-_ has joined #haiku
tuaris has joined #haiku
zardshard has left #haiku [Error from remote client]