<nekobot>
• OscarL (f164c0c9): bitstring: drop outdated and unused recipe. (#11926)…
<howard_[m]>
Good afternoon everybody 😄 i am currently trying to solve the easy ticket using wsl2+ qemu
<howard_[m]>
But can anyone know is there a optimize workflow for contributing code
<howard_[m]>
Like a faster way maybe
<Begasus[m]>
Hello howard_ that's not my cup of tea, maybe some of the real dev's can answer if they are around
wicknix_ has joined #haiku
wicknix has quit [Ping timeout: 480 seconds]
OscarL has joined #haiku
<OscarL>
Hey there Begasus[m]. Thanks for the speedy merges :-)
tulip[m] has left #haiku [User left]
<OscarL>
wrote some bash script to sort the outdated dev-python/*/*.recipes in order of most to less used/requiered... there were 23 with 0 usage on-tree... so I spent some hours reading git logs to see if we could just drop them instead of updating them :-)
<Begasus[m]>
Hi OscarL !
<OscarL>
only 2 of those 23 "survived" ("pyserial" because it has a Haiku specific patch, and "rst2pdf" because it provides a useful "cmd:", requires a "non any" python package, so having it package can be convenient).
<Begasus[m]>
thanks there, I guess most of them are fine with pip now, only left one open (want to give auggiedogie a change to report on that)
<OscarL>
Ideally, we should drop most things that are "any", without patches, with no non-any dependencies, etc. Some stuff is just not that useful any longer.
<Begasus[m]>
down to 17 now :P
<OscarL>
more and more stuff will start to move away from setuptools also (into build+installer).
<Begasus[m]>
after cocobean's PR for libkmahjongg it stroke me that I got tons of packages already moved to KF6 ... so trying to do some catching up there :)
<Begasus[m]>
that's good OscarL !
<Begasus[m]>
easier to get around with those
<OscarL>
re: kf6... can imagine. easy to lose track of things once you're more than +10 package deep into recipe changes :-D
<OscarL>
using "git log" more and more... /me starts to hate those "bump version" commit messages that do not specify the version on the summary/subject line :-)
<Begasus[m]>
you want to see something like "bump version to 24.12.3" then?
<OscarL>
makes "git log --oneline" pretty useless, if you need to look into the commit itself to see what version changed.
<OscarL>
Begasus[m]: that would be preferable, yes. I have been doing that (almost) always.
<Begasus[m]>
OK, will try to keep that in mind then :)
<OscarL>
another one that would be nice... when adding a new recipe, to mention if it is a requirement for some other package, or not.
<Begasus[m]>
I'm still facing some "KIO" issues, so apps using that will remain on both frameworks for now
<OscarL>
thanks Begasus[m]. Now I need to convince korli to also mention versions on bumps :-P
<Begasus[m]>
tried to do that with those 2 new ones yesterday
<OscarL>
good. after some time passes... it gets hard to remember details, so always nice if that info can be found on the logs.
<OscarL>
(even if "git log -S" can be SO SLOOOOW :-/)
<Begasus[m]>
it makes sense for me to just mention "bump version" on most of them, but yeah, could be handy for those tracking down the logs
<OscarL>
adding that "--since=YYYY-MM" does helps, though.
<Begasus[m]>
had a few scripts to grab commits for like the last month/quarter
<OscarL>
today I spent many hourse on git log. makes you appreciate well written commit messages :-)
<Begasus[m]>
used those for the haikuports reports, should set it up again, just can't find them anymore :P
<Begasus[m]>
+1
<OscarL>
I wrote some quick notes about packages under dev-python. Couple of quick "Do's and Do Not's". Might be good to add them to the wiki, once I polish them a bit.
<OscarL>
Example... We seem to have many packages that require "cmd:python3", but also depend on some specific python package, say: "setuptools_python310". That's wrong, and the recipe should depend on "cmd:python3.10" on that case.
<OscarL>
otherwise, if we change the default Python, we're breaking all those recipes.
<Begasus[m]>
well, we have python3 to be the default here
<Begasus[m]>
err ... you would have to change every recipe that uses the default from 3.10 to 3.* then?
<OscarL>
some packages may legitimally only depend on cmd:python3 (because they don't care about which version, and do not have any requirement on external packages). That's OK.
<Begasus[m]>
maybe some bashism to get the version and fill that in to a variable?
<Begasus[m]>
cmd:python3 is only provided by python3.10 just for this?
<OscarL>
yes, using "$pythonVersion" or similar should be enough.
<Begasus[m]>
just like with lua?
<Begasus[m]>
bugger there's that l word again :P
<OscarL>
But think what would happen if (when) we make 3.12 the default?
<Begasus[m]>
well, you could start ahead and check how it works for 3.10? :P
<OscarL>
not being able to change the default python at will complicates testing.
<OscarL>
thus why I put that draft PR for consideration.
<Begasus[m]>
will only be an issue if you got multiple python versions involved like "PYTHON_PACKAGES=(python39 python310 python311 python312)
<Begasus[m]>
"
<OscarL>
and looking at the logs... waddlesplash pondered/suggested having "_default" packages for Pythons, as done for OpenJDK, when he changed the default from 3.7 to 3.9
Nasina has joined #haiku
<Begasus[m]>
whoot, that must have been some time ago :P
<OscarL>
no, the issue is the version mismatch between whatever cmd:python3 points at, and recipes asking for, say, _python310, or _python39, etc version of a given package.
<OscarL>
Begasus[m]: told you I spent some hours digging at logs :-P
<Begasus[m]>
heh
<OscarL>
best I could do during the day, without internet :-D
<Begasus[m]>
there's a good in the bad :)
<OscarL>
Been trying to polish my script that grabs data from PyPi.org too. Works well enough now... but still need to test if I can make it only download data for packages that have new versions since last checked.
AlienSoldier has quit [Quit: Vision[]: i've been blurred!]
<OscarL>
Will probably run that in a few minutes, and probably kill the few remaining MBs I have availabe this night, lol :-D
<OscarL>
Begasus[m]: good thing I had a copy of bash html docs. Man... the more bash I learn, the more crazy I feel :-D
<Begasus[m]>
even more? lol
<OscarL>
That was my reaction as well. Go figure!
<Begasus[m]>
well ... I'm nitpicking on my own recipes :)
<OscarL>
same here with the style of "my" python (packages) recipes. Keep finding small tweaks to make it easier to read.
<Begasus[m]>
gwenview needs the haiku plugin for the tests ...
<Begasus[m]>
hate those tests the appear to be idle ... you never really know if they are actually doing something :P
<OscarL>
on the other hand, at least they don't make you waste much power, compared to those test that seemingly hang at 100%
<Begasus[m]>
also true :)
<OscarL>
and you never know if they are close to finish, or just hang with hight cpu usage :-/
<OscarL>
will run my "pypi-versions.py" script to see if just updates stuff as intended. I might just disconnect in panic otherwise.
* OscarL
waves pre-emptively
<Begasus[m]>
OK, see you around!
<OscarL>
mmm, seems it is just downloading all the json data again. Not sure if bug on my script, or I just don't understaind this etag header thing is supposed to work. pypi.org is not sending any 304 (not modified) return code :-(
<Begasus[m]>
eeps
<OscarL>
"[164 of 219] - pythran" and started to have connection issues. I wonder if I'm beeing throttled (I have sleep() calls between each request(), and even a 5 seconds pause each 50 packages, just in case).
<Begasus[m]>
almost there then
<OscarL>
darn aiohttp.picle is 5.09 MiB. What the hell the yare returning in that .json ?!!? :-D
<OscarL>
s/the yare/they are/
<Begasus[m]>
you got the rest of the day to figure that out? ;)
<OscarL>
probably :-D
<OscarL>
but first... some bed time.
<Begasus[m]>
k, sleep well
<OscarL>
Nice to read you Begasus[m], as always. Have a fun day!
<nekobot>
• Begasus (695ad33c): libkdcraw_kf6, bump to 24.12.3 (#11947)
Nasina has quit [Read error: Connection reset by peer]
Nasina has joined #haiku
Nasina has quit [Read error: Connection reset by peer]
Nasina has joined #haiku
Nasina has quit [Read error: Connection reset by peer]
Nasina has joined #haiku
freddietilley has quit [Ping timeout: 480 seconds]
freddietilley has joined #haiku
Nasina has quit [Read error: Connection reset by peer]
Nasina has joined #haiku
mmu_man has joined #haiku
nephele has joined #haiku
<nephele>
howard_[m]: A faster way than what?
Nasina has quit [Read error: Connection reset by peer]
<nephele>
zardshard, waddlesplash: around?
Nasina has joined #haiku
Nasina has quit [Read error: Connection reset by peer]
freddietilley has quit [Quit: WeeChat 4.5.0]
freddietilley has joined #haiku
Nasina has joined #haiku
Nasina has quit [Read error: Connection reset by peer]
Nasina has joined #haiku
Nasina has quit [Read error: Connection reset by peer]
maylay_ has joined #haiku
maylay has quit [Ping timeout: 480 seconds]
<waddlesplash>
nephele: briefly, not for long
HaikuUser has joined #haiku
HaikuUser is now known as wire_panel
<nephele>
waddlesplash: Regarding the webkit keyboard patch on the github... PulkoMandy asked about the lifetime of the nativeevent, and suggested possibly using a copy of the BMessage as a pointer. That didn't really work out (because webkit wants to pass a pointer back to it later)
<nephele>
Can you see/check the lifetime of that object and explain to me when it has to be deleted?
<waddlesplash>
I don't know anything about how these objects work
<waddlesplash>
what prevents you from doing the analysis yourself?
<nephele>
Pretty much No Haiku code I've seen manually deletes pointers. So don't know that much about it. I know how it works "in general" but in C++ specifics It still really confuses me
<waddlesplash>
not sure what you mean by "no Haiku code", we do it all over the place
<nephele>
I wrote what I ment. "No Haiku code I've seen"
<waddlesplash>
in internals, that is. in "user" code it's less common but that's because BMessage and BLooper have their own management
<waddlesplash>
well, then all you are saying is that you have not seen much Haiku code
<nephele>
Okay? Why do you want to argue about this?
<waddlesplash>
well it's a bit surprising to me is all, because even if you just read commit logs as they went by over the past year, you would have seen a lot of it
<nephele>
I don't follow your analysis, and I don't see how that is relevant. Of course I've seen pointers be deleted. But that doesn't mean that I understand the full lifetime management required for it
<waddlesplash>
anyway not sure what there is that's confusing here. You get memory with new/malloc, and when you're done with it, you release it with delete/free. You can't release it while someone else could still access it later
<waddlesplash>
The "lifetime management" is entirely up to the application that called new and will later call delete
<waddlesplash>
if I have a class that calls new, and never "gives out" the pointer to that memory, then the "lifetime management" is quite easy: it can delete this at any time because nobody has any references to it besides itself
<nephele>
I don't see any place where the message for this is created with new or malloc
<waddlesplash>
then that's a good sign you are not allowed to delete it
<nephele>
Then I don't understand the code review i got or it :/
<waddlesplash>
unless the documentation (any documentation) says "you own the pointer returned from/passed into this function", you generally don't, so you can't delete it
<wire_panel>
Hello, I am new to Haiku, does the dev branch just roll up consistently?
<waddlesplash>
nephele: your original message here said "using a copy of the BMessage". that means allocating a new one with "new" that is a copy of the old, which then you would of course fully own and can do whatever with
<nephele>
wire_panel: the nightly(master) branch is in active dev. the r1betaX branches are "stable", and only the last one of them receives bugfixes
<waddlesplash>
wire_panel: not sure what you mean? the dev branch is built into nightly builds once a day
<nephele>
waddlesplash: yes sure, I can copy it from wherever Webkit gets the native event from. PulkoMandy suggested this as an option to prevent a memory leak that *could* happen from the passed in pointer if it is not deleted
<nephele>
But I have no clue where the original message is created (maybe in minibrowser when it sends the BMessage?) or where I'd have to clean up the pointer
<waddlesplash>
well I don't know where it's created either, you will have to investigate
<nephele>
Hmm. I guess I'll leave the code as is, with the native event beeing a pointer. If there is a possible memory leak that can be pointed out in code review again...
<waddlesplash>
if PulkoMandy says there is possibly a bug, that will need to be determined
<waddlesplash>
investigating what other platform code does here will probably help you out
<nephele>
Other platforms use smart pointers
<waddlesplash>
then why not also use a smart pointer?
<nephele>
They all use webkit specific smart pointer types. I don't know how those work yet. :)
<nephele>
and I should know what problem I am solving first...
<waddlesplash>
time to read the docs
<nephele>
if you have any cool webkit docs feel free to share them...
<waddlesplash>
generally in large codebases like WebKit it's best to just copy what everyone else is doing and not question it too much, unless you really have some specific problem to solve
wire_panel has quit [Quit: Vision[]: i've been blurred!]
<waddlesplash>
so, try changing to a RemainPtr and see what happens
<nephele>
so RetainPtr<BMessage> m_nativeEvent;?
<waddlesplash>
ah, looking more closely, this is only used for Apple platforms
<nephele>
nope. sais incomplete type... hmm
<waddlesplash>
so that won't work here
<nephele>
yeah, but every port seems to have their own unique pointer type here
<nephele>
so... I don't know which port to just copy it from :P
<waddlesplash>
did you look at the other implementations at all?
<waddlesplash>
I just glanced at GUniquePtr, it's an alias for std::unique_ptr in the general case
<waddlesplash>
so, try using that then
Nasina has joined #haiku
<nephele>
Not for the pointers. I was mostly still looking at keyboard stuff
<nephele>
why not use std::unique_ptr directly though?
<waddlesplash>
it looks like there may be specializations for specific types
<waddlesplash>
glib-specific types
<nephele>
Okay. But for us that shouldn't matter
<phschafft>
Guten Morgen.
<nephele>
Hello phschafft
<nephele>
I don't understand how the unique ptr works. If i use it i have to adjust all code that wants to use the pointer... But don't know to what
<waddlesplash>
it's a standard C++ feature, go read the docs on it
<nephele>
are there good docs that aren't on the cppreference site?
<waddlesplash>
what's wrong with cppreference?
<nephele>
No dark mode
<waddlesplash>
so use reader mode?
<nephele>
How am I supposed to use reader mode? Have you implemented it recently...?
<waddlesplash>
no, use a browser that has it
<nephele>
I don't want to install another OS on my computer. I'm developing on Haiku not on linux
<waddlesplash>
we have browsers that have it on Haiku
Nasina has quit [Read error: Connection reset by peer]
<nephele>
I doubt it. the ported browsers don't even support the dark mode
<nephele>
Yeah. I'm not installing mozilla adware that doesn't even support my keyboard...
<waddlesplash>
this (theoretically) doesn't have mozilla telemetry
<waddlesplash>
and I don't know why it wouldn't support your keyboard, but if it doesn't, that's a bug in the Wayland layer's keymap handling and could be fixed
<nephele>
Nobody patched any defaults, and the only problem was a failed port of the telemetry. I don't really trust that.
<waddlesplash>
then learn how to use about:config and/or a hosts file and turn it all off
<waddlesplash>
not too hard
<nephele>
No.
<waddlesplash>
anyway, if you are going to be arbitrary and look for excuses rather than solutions, I am under no obligation to help you
<nephele>
Sure, you are more looking to be right in an argument that anything else...
<nephele>
I don't care what junk people port. I use Haiku precisely to not have to deal with user-hostile software
<waddlesplash>
no. you asked for a page with documentation, but then said "not cppreference because no dark mode", I said "use a browser with dark mode", you said "we don't have any on Haiku", I proved that false, and it turns out you will refuse to use that browser for reasons that have nothing to do with dark mode
<waddlesplash>
that's your choice, but I am also under no obligation to assist you with that choice
<nephele>
also fyi: the reader mode is the wrong tool for this. System colors support is the correct tool
<nephele>
No, you said reader mode. That's a different thing and only works with sites that have explicit support for it
<waddlesplash>
wrong
<nephele>
" <waddlesplash> so use reader mode?"
<waddlesplash>
reader mode it works on all webpages that have main content clearly discernible from the rest of the page in the markup
<waddlesplash>
which should be all pages if they follow accessibility standards correctly
<nephele>
They don't. Tough luck.
<waddlesplash>
well, this page seems OK in reader mode
<waddlesplash>
some stuff is not in monospace which should be, a bit annoying but still entirely readable
<waddlesplash>
I agree, system colors support on the page would be nice, but this workaround works for a page I would look at for 5 minutes
<nephele>
I'm not talking about the page using system colors
<Anarchos>
@nephele i have no haiku machine up right now, but i wonder if Falkon has a dark mode
<nephele>
I'm talking about the forced system colors css way where the User agent ignores color directives
<nephele>
Anarchos: WebPositive supports dark mode fine. But no forced colors
<waddlesplash>
sure, and that would also be nice (and Gecko supported that too last I looked)
<nephele>
maybe firefox has a way to use this properly in about:config. But I'm not going to find out
<waddlesplash>
I just linked a support page on Firefox that details an option "override the colors specified by the page"
<waddlesplash>
so I am replying "no" to your thought
<waddlesplash>
"I thought X" -> "No, X is incorrect, it's really Y"
<waddlesplash>
this is an entirely standard conversational pattern in English
<nephele>
Maybe you should have written that, if you wanted to say that
<nephele>
No it isn't
<waddlesplash>
I am a native English speaker who lives in an English-speaking country
<waddlesplash>
so I would think I have at least some authority to attest to what is in common usage lol
<nephele>
Statistically, that makes you less capable of english speaking than foreigners who learn the language. So I don't know why you keep bringing this up
<nephele>
and it's hardly a defense. "Appeal to authority", in this case yourself...
<nephele>
and it does not matter. you are not in a chatroom of native english speakers. It's not nice to behave in a rude way and then claim that's just the way it's supposed to be.
<waddlesplash>
I have an experience that this pattern is in common usage. This pattern is supported by the fact that I have lived my whole life speaking English. Sure, it's a limited sample size, but it's still a lot bigger than 1
<waddlesplash>
this pattern is not rude at all in English
<nephele>
It's not rude *wherever you live*
<waddlesplash>
there are plenty of patterns in any language which would be rude if literally translated into some other language
<nephele>
Claiming that you have a significant sample of english is ridicilous. There are many more english speaking countries that have other norms of what is and isn't rude
<waddlesplash>
there are German idioms I've seen translated into English which would be rude if said literally
<waddlesplash>
(but if reworded/rephrased can be made to mean the same thing and not be rude)
<waddlesplash>
this is just how languages work
<nephele>
Sure? and? that doesn't mean I'd instult someone and then say "deal with it that's just the language"
<nephele>
Then do this rephrasing. "No" does not mean the same as "I disagree" or "you are incorrect"
<waddlesplash>
well, you are incorrect, because it quite often means the latter
<waddlesplash>
it's a simple negation, that's it
<nephele>
Yes a negation towards a mental model. That is rude. It is the same as someone saying "I saw you walk by" and you respond "No you didn't"
<nephele>
instead of "I did not walk by"
<Anarchos>
waddlesplash: could you find time to look at my pull requests ? They are not important anyway, i am just curious.
<waddlesplash>
nephele: People do say "no you didn't" to "I saw you walk by" lol, that's perfectly common in English
<waddlesplash>
"I saw you walk by this morning"; "No, I didn't walk by, that must've been someone else"
<waddlesplash>
quite typical exchange
<waddlesplash>
wouldn't be seen as rude at all
<nephele>
You produced my second example, that exchange is not rude.
<waddlesplash>
but it starts with "no"
<waddlesplash>
what is being negated here, except the "I saw you"?
<nephele>
Yes, that isn't the problem. The problem is you using no without any qualification
<waddlesplash>
"No, you didn't, I wasn't out this morning"
<waddlesplash>
that is also fine
<waddlesplash>
Anarchos: every time I touch bootloader code it seems to get really complicated very quickly lol
<waddlesplash>
I did look at your patchset before, what happened to my review?
<nephele>
What is the problem with the bootloader args btw? don't we have a version field for that? It seems every time this comes up someone sais we can't break compat...
<Anarchos>
waddlesplash: nobody reviewed them after you so they are still active...
<waddlesplash>
Anarchos: if you mean https://0x0.st/8XcT.report it looks like an invalid pointer being referenced
<waddlesplash>
try with the guarded heap and see what happens
<waddlesplash>
nephele: we do have a version field, but we still can't handle changes in the middle of the structures without breaking compat because they shift all fields downward. The version field mostly just ensures the kernel will not read garbage data
<waddlesplash>
and will error out instead
<Anarchos>
waddlesplash: you read in my t houghts :) i will try, but if i remember Debugger didn't give clues. Anyway i will test that tonight again.
<nephele>
waddlesplash: doesn't the bootloader know which kernel it boots?
<waddlesplash>
good question, I don't know
<waddlesplash>
I've never tried playing with this too much
<waddlesplash>
if it doesn't that sounds like something we should fix
<nephele>
If it doesn't, it should learn this in the future. Even if it can only detect "new" kernels then. It can use the old args for older kernels, and a new arg model for newer ones
<waddlesplash>
the problem is usually the other way around
<waddlesplash>
an older bootloader and a newer kernel
<waddlesplash>
converting the structures is not easy unfortunately
<waddlesplash>
a much better idea here would be to make the EFI bootloader capable of updating itself
<waddlesplash>
"Too new kernel. Update bootloader?" or something like that
<waddlesplash>
and then it updates itself inside the ESP on the fly
<nephele>
I don't like that idea. though it is possible
<nephele>
I'd rather the efi loader be updated on kernel updates aswell
<waddlesplash>
I wouldn't
<nephele>
but again, this requires efi var supports
<nephele>
... why?
<waddlesplash>
I expect my ESP will not change without something asking me first
<waddlesplash>
and I don't want to update the EFI loader in there except rarely for a variety of reasons
<nephele>
That's a wierd assumption. I am not talking about the case of you manually installing the loader and installing that
<nephele>
I am talking about Haikus installer having installed this itself, and managing it going forward
<waddlesplash>
not that weird, it's a separate partition, if I don't mount it then it won't change
<nephele>
It's an implementation detail. Why do you want it to *not* update the efi loader?
<waddlesplash>
sure, that might be the one case it makes sense
<nephele>
it seems super wierd that it can't be updated automatically, but has an updater feature build in to complain
<waddlesplash>
for the same reason I don't like Windows or GRUB changing boot order automatically?
<waddlesplash>
it's the exact same thing
<nephele>
It's not changing the boot order
<waddlesplash>
nephele: no, the updater feature would be able to update it
<waddlesplash>
but from the bootloader itself
<nephele>
This makes our update process two step for efi pcs... not nice
<nephele>
and also uneccesary
<waddlesplash>
like I said: the "Haiku Installer generated the ESP" case is the one where it does make sense to auto-update
<waddlesplash>
all other cases, no
<nephele>
Yes. For all other cases this seems so niche that I don't see why we would have to make a complicated updater
<waddlesplash>
also, keep in mind some people may use custom Secure Boot keys, so you can't auto-update the bootloader because it'll break booting
<nephele>
That's not relevant. and wouldn't work with your idea either
<waddlesplash>
no it wouldn't, but it would mean that every time it would happen the user gets an easy chance to say "no"
<nephele>
updating in-place is a bad idea. you might even destroy your bootloader in this update, and instead of beeing able to start an older haiku to fix it you are now locked out completely
<waddlesplash>
so copy the file and leave the old bootloader around with another name
<nephele>
that's what the automatic update would do
<nephele>
when installing a new haiku version
<waddlesplash>
right now this would happen every nightly update though, which will result in dozens or more bootloaders
<nephele>
also, if you copy it, and you are still in the loader. this implies that the efi loader now also has to expose a ui to copy files or we have to ship the edk2 shell
<waddlesplash>
?
<nephele>
It shouldn't. It should only update if the image is different
<nephele>
> <waddlesplash> so copy the file and leave the old bootloader around with another name
<waddlesplash>
the image will basically always be different
<nekobot>
[haiku/haiku] 80f796971464 - kernel/x86_64: Correct address mask in DebugGetReverseMappingInfo.
<nephele>
Still doesn't matter. You can a) ignore the update if it's just the hrev) or b) limit the number of loaders
<nephele>
anyhow, the automatic update should be added atleast for the betas
<nephele>
for the nightlies a network connected efi loader seems way too huge in scope for such an issue
<waddlesplash>
network connected? what?
<waddlesplash>
we ship the EFI loader inside the system packages
<waddlesplash>
it should update itself from there
<nephele>
that's be less complex. but still a bit overkill
<nephele>
either you need a in-place update, or you need to add facilities to tell users where to place the old loader or new loader *and* efi runtime services support
<nephele>
the cpp reference is very unhelpfull for the std::unique_ptr
<waddlesplash>
well there are other tutorials online, take a look around
<waddlesplash>
bbl
<nephele>
maybe. I didn't see anything that deals with classes passing this around later, and their return types *shrug*
erysdren has joined #haiku
HaikuUser has joined #haiku
zardshard has left #haiku [Disconnected: Replaced by new connection]
zardshard has joined #haiku
bbjimmy has quit [Quit: Vision[]: i've been blurred!]
bbjimmy has joined #haiku
yann64 has joined #haiku
yann64 has quit [Quit: Vision[]: i've been blurred!]
<dovsienko>
well, that's exactly 2496KiB, in case anyone remembers the significance of this number
<Skipp_OSX>
maybe debug in boot loader tries to build a floppy image?
<Anarchos>
Skipp_OSX i didn't put the debug flag on "boot loader", on purpose
HaikuUser has joined #haiku
<Anarchos>
Skipp_OSX ah i see some binaries left in debug_1/system/boot ....
Nasina has quit [Ping timeout: 480 seconds]
HaikuUser has quit []
<Anarchos>
i was sure to have launched "jam clean"…
neoncortex has joined #haiku
<neoncortex>
hey, I have a doubt: if I apply a bunch of attributes to files, and directories, how do I backup it?
<neoncortex>
sorry if that sounds noobish, I'm just starting with haiku, so-
<Anarchos>
neoncortex what do you mean by 'backup' ?
<neoncortex>
Anarchos: say I transfer that files to another partition. I will lose the attributes?
<Anarchos>
neoncortex if it is a bfs partition, i would say you keep them (just try). If it is on a non attributes partiition (FAT or ext2...) i am pretty sure you loose them
<Anarchos>
neoncortex but as i said, you can easily try it
<neoncortex>
sure.
<jmairboeck>
neoncortex: you can also zip them and backup the archive. That should preserve the attributes.
<neoncortex>
so I can zip a directory, and all the files, and directories, will hold the attributes. That solves it.
Anarchos has quit [Quit: Vision[]: i've been blurred!]
Nasina has joined #haiku
coronavirus has joined #haiku
neoncortex has quit [Ping timeout: 480 seconds]
Nasina has quit [Read error: Connection reset by peer]