<nephele_xmpp>
I’ve just ordered two. I can pay by bank transfer, hooray!
<Hiryu>
I may get 1 and order another if I like it... or just get 2 from the start
<nephele_xmpp>
I now hope it arrives this month, then xD
<nephele_xmpp>
fosdem is start of next month
<Hiryu>
Too bad my father is gone... he would have appreciated the KDE mug
<Hiryu>
I think I'll get a Haiku and a KDE mug
<nephele_xmpp>
I ordered a pol shirt and two mugs now
<nephele_xmpp>
I hope it works out till fosdem, who knows :D
<nephele_xmpp>
I hope you are happy with your KDE and Haiku mug aswell
<nephele_xmpp>
I also use KDE sometimes, mostly on the steam deck
<moparisthebest>
Anyone ran haiku on the steam deck yet?
<nephele_xmpp>
Yes, It works on the beta release but not on the nightlies
<nephele_xmpp>
screen is rotated wrongly, we don’t support it’s input devices or video over usb-c at the moment however
<moparisthebest>
Neat!
x10z has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
<Anarchos>
Strange, now that i can get bios drive id for all the drives, BootManager is unable to collect their partitions :)
<nephele_xmpp>
Anarchos: I commented on your patch where the build issue came from
<Anarchos>
nephele_xmpp ah thanks i will look
flag has quit [Ping timeout: 480 seconds]
mmu_man has quit [Ping timeout: 480 seconds]
ablyss has quit [Quit: Konversation terminated!]
dru_satori has quit [Quit: Vision[]: i've been blurred!]
dru_satori has joined #haiku
<dru_satori>
so, I might be an idiot ( totally probable ), but. how do I update a commit to gerrit? a --amend does not include the additional file change in response to a comment. ( yeah yeah, stupid question I am certain, but I've had a beer or four )
<dru_satori>
hoping I did not screw this up ( too many years doing things where I am 99% the only commiter )
<Skipp_OSX>
no it's not a stupid question, amend is right, make sure that the Change-Id at the bottom is the same as on Gerrit. It looks like you already pushed a new change so you'll have to abandon the old one now.
<Skipp_OSX>
The Change-Id is what identifies the commit as the one to update on Gerrit.
<dru_satori>
I think they are both wrong, because neither is picking up the additional file that was altered.
<Skipp_OSX>
well abandon one and update the other then, just make sure in your new commit the Change-Id is identical.
<dru_satori>
as soon as I figure out how, I will :D
<dru_satori>
so, I thought I had them the same, but apparently not
<dru_satori>
I'm following the 101 on the Haiku dev.
<dru_satori>
installed the hooks.
smalltalkman__ has joined #haiku
<dru_satori>
but it looks like I got 3 change id's
<Anarchos>
dru_satori you must edit the commit message and change the Change-Id
<Skipp_OSX>
no worries
<dru_satori>
I see that now, I think I got it sorted now
<dru_satori>
abandonded the bad ones
<dru_satori>
may have lost Nephelle's comments, but the end result is the desired one I think
<dru_satori>
nothing like learning new tools
chaosDNA has joined #haiku
<dru_satori>
I missed a step in the middle. lesson learned, notes taken and saved
chaosDNA has quit []
chaosDNA has joined #haiku
dru_satori has quit [Quit: Vision[]: i've been blurred!]
chaosDNA has quit []
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<nephele_xmpp>
dru_satori: you have to use git add to mark the files to be commited as before, and then use git commit —amend
<nephele_xmpp>
skipp_osx: I gotta say though, your changesets (the biigger ones) are really hard to review because it has 100s of revisions by now that usually don’t change anything
<nephele_xmpp>
and it’s hard to figure out which change needs whiich other one, gerrit knows but doesn’t really tell… annoyingly so
<Skipp_OSX>
sure but I do, in the comments
<Skipp_OSX>
The Tracker Shortcuts PS has gotten out of hand.
<Skipp_OSX>
I tried to split it out, but so much is interconnected it is tough to do.
<nephele_xmpp>
Let’s try to get this patch series sorted. maybe next time we can work with a feature branch in gerrit for such a bigger refactor
xet7 has quit [Remote host closed the connection]
<Skipp_OSX>
I can't promise this is the last one...
<Skipp_OSX>
Tracker needs more refactoring... a lot more
<nephele_xmpp>
I know :)
linuxmaster has joined #haiku
Shaka444 has joined #haiku
vdamewood has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
duncsauce has quit [Remote host closed the connection]
duncsauce has joined #haiku
duncsauce has quit [Ping timeout: 480 seconds]
duncsauce has joined #haiku
duncsauce has quit [Ping timeout: 480 seconds]
duncsauce has joined #haiku
duncsauce has quit [Ping timeout: 480 seconds]
duncsauce has joined #haiku
duncsauce has quit [Read error: Connection reset by peer]
duncsauce has joined #haiku
duncsauce has quit [Ping timeout: 480 seconds]
duncsauce has joined #haiku
x10z has joined #haiku
<Halian>
o/
<Begasus[m]>
\o
<Begasus_32>
back to 64bit
Begasus_32 has quit [Quit: Vision[]: i've been blurred!]
Begasus has joined #haiku
duncsauce has quit [Ping timeout: 480 seconds]
<phschafft>
:)
<Begasus[m]>
Hi there phschafft :)
duncsauce has joined #haiku
x10z has quit [Ping timeout: 480 seconds]
<phschafft>
all good?
<Begasus>
not bad, so good enough :)
<phschafft>
:)
<Begasus>
how's it going there?
duncsauce has quit [Ping timeout: 480 seconds]
<Begasus>
build completed for haikuwebkit on 32bit PulkoMandy :) +1
<phschafft>
hm, just woke up.
* Begasus
hands a morning cup of coffee and a chocolate for phschafft
<Begasus>
that should get you started :)
<phschafft>
thanks for the chocolate. but coffee is not my style. :)
<Begasus>
will keep that in mind :)
<andreasdr[m]>
Hi there.
<Begasus[m]>
Hi andreasdr coffee for you then?
<phschafft>
was dreaming about a text compression thingy.
freddietilley has quit [Ping timeout: 480 seconds]
<Begasus>
got it fixed in your dream?
freddietilley has joined #haiku
<phschafft>
the bit structure is solid. also it seems that it could actually compress text a bit. currently thinking about the best way to do streaming mode encoding.
<phschafft>
which, as always with streaming compression is: you could do so much better if you would know the future! ;)
<phschafft>
but I guess, that is not just about compression ;)
freakazoid332 has joined #haiku
<Begasus>
still stuck in the past here (guess that's the age) :)
frkzoid has quit [Ping timeout: 480 seconds]
<phschafft>
;)
MisthaLu has joined #haiku
HaikuUser has joined #haiku
x10z has joined #haiku
bjorkintosh has quit [Ping timeout: 480 seconds]
duncsauce has joined #haiku
mmu_man has joined #haiku
duncsauce has quit [Ping timeout: 480 seconds]
x10z has quit [Ping timeout: 480 seconds]
Nasina has quit [Read error: Connection reset by peer]
Nasina has joined #haiku
freddietilley has quit [Ping timeout: 480 seconds]
<phschafft>
decoder is written, now the encoder. that is the hard part.
freddietilley has joined #haiku
<Begasus>
basics done :) grabbing luminance_hdr-2.6.0-1-x86_64.hpkg and moving it to /Share/haikuports/packages/luminance_hdr-2.6.0-1-x86_64.hpkg
freddietilley has quit [Ping timeout: 480 seconds]
Nasina has quit [Read error: Connection reset by peer]
freddietilley has quit [Ping timeout: 480 seconds]
<phschafft>
Begasus: so far it compresses the test text by 0.8%! ;)
freddietilley has joined #haiku
flag has quit [Ping timeout: 480 seconds]
<Begasus>
Hello Anarchos
x10z has joined #haiku
<Begasus>
how do you compress text phschafft? :) (remove the spaces?) :D
<Begasus>
do we have ICC_PATH ?
<Begasus>
we do :)
<phschafft>
basically ;)
<phschafft>
this thing is designed to work with *very* low amounts of RAM. so the compression is not expected to be good.
<Begasus>
considering the RAM, 0.8 isn't that bad I guess?
<phschafft>
the idea is that it has four modes of operation: for ASCII characters encode-as-is, 3-byte-to-4-byte binary mode, run length coding (1 byte raw + 1 byte every 32 extra copies), and a 2 byte sequence replace with 32 dictionary entries.
<phschafft>
and I think there is a lot of things you can put into the encoder to will change the compression like a lot.
* Begasus
is lost after 0.8% :D
BrunoSpr has quit [Quit: Vision[]: Ich wurde eingeweicht!]
BrunoSpr has joined #haiku
x10z has quit [Ping timeout: 480 seconds]
<phschafft>
;)
<Begasus>
stupid LuminanceHDR with hardcoded paths
<Begasus>
opens fine, but can't find it's own documentation
HaikuUser2 has joined #haiku
HaikuUser2 has quit []
duncsauce has joined #haiku
<phschafft>
Begasus: basically there are two modes of storing bytes uncompressed (so you can store data even if you can't find a compressable pattern). and two modes of storing them compressed (one is to store any byte that repreats up to 33 times as 2 bytes) and the other is to store any combination of two bytes as one byte.
<phschafft>
the two mores that are for strong uncompressed data are in a way that worst case ASCII text is returned as is (so no win no loose), and binary data is 1/4 bigger.
x10z has joined #haiku
<phschafft>
and the best case is something like going down to 1/32, e.g. for a stream of 0-bytes.
duncsauce has quit [Remote host closed the connection]
freddietilley has quit [Ping timeout: 480 seconds]
<BrunoSpr>
hello all, hello Begasus
duncsauce has joined #haiku
freddietilley has joined #haiku
<Begasus>
Hi BrunoSpr
<BrunoSpr>
Begasus, I see you compiled LuminanceHDR, I tried to compile but no good result here, at least I learned a bit
<Begasus>
there are some libraries involved BrunoSpr :)
<BrunoSpr>
Begasus, weekend I will try again if I take some time then
<Begasus>
for packaging at least data directory needs to be changed in the source
<BrunoSpr>
Ah, thats why... anyway to tired atm
<Begasus>
np, cu l8r :)
<Begasus>
afk for a bit too here
<BrunoSpr>
Sounds too difficult for me, so they need a bugreport then?
<BrunoSpr>
If it is too complicated, better to postphone then for now?
zard has joined #haiku
flag has joined #haiku
x10z has quit [Ping timeout: 480 seconds]
Anarchos has quit [Remote host closed the connection]
<geniodiabolico>
I understand the wanting to switch to BString operations, but this whole "It shouldn't be larger than this size" doesn't make sense to me. The supertype and subtype exist and can be up to a size. Combining both into a buffer of the same size as each means that if they are greater than 1/2 that size you truncate. That guarantees that the subsequent lookup will fail and the whole thing is useless.
<nephele_xmpp>
geniodiabolico: a mime type longer than B_PATH_NAME_LENGTH can’t exist on our system, because the mimeDB stores the mime names as files on the filesystem
<nephele_xmpp>
So, use BString to fix the truncation issue. But the path name length makes no sense to remove because this can’t be supported with how the system currently works
<Begasus>
next round, install to appsDir/LuminanceHDR
x10z has joined #haiku
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
MisthaLu has quit [Quit: Leaving]
x10z has quit [Ping timeout: 480 seconds]
<phschafft>
Begasus: got to -7.977%!
MsInput has joined #haiku
<phschafft>
but it depends on the input. for some files I get better results. for some worse, naturally.
<Begasus[m]>
don't overdo it phschafft :D
<phschafft>
;)
<phschafft>
the decoder only holds at max 65.5 bytes of state.
<phschafft>
so the total RAM usage of the decoder will be somewhere around 100 byte.
<Begasus>
k, help inline working now :)
<Begasus>
talking bytes ... you don't see that a lot these days :D
<Begasus>
cli still seems to do it's work
Anarchos has joined #haiku
<phschafft>
hm?
<phschafft>
ok, I reduced the number of dictionary slots 32 -> 16, now it's -10.214%. so... I guess my logic could still do better encoder side. ;)
dru_satori has quit [Quit: Vision[]: i've been blurred!]
<phschafft>
(over a number of test files it's not a clear picture if more slots are better, hinting towards more encoder work :)
<phschafft>
however that would reduce the decoder memory to 33.5 bytes.
<phschafft>
(2 bytes per slot + 1.5 bytes for other state)
dru_satori has joined #haiku
yann64 has quit [Quit: Vision[]: i've been blurred!]
<nephele_xmpp>
Uhh, how does Jam get build for the haiku crosstools?
<nephele_xmpp>
I’m building it on macos now, got the cross tools build but i don’t have jam, and am somehow failing to find the docs where this is installed/compiled
<nephele_xmpp>
ah, found it
<nephele_xmpp>
it’s in the pre-quisites *above* all the other instructions
<Begasus>
first step, then run? ;)
<nephele_xmpp>
begasus: i checked the macos section because I am on macos :D
<Begasus>
time to update there then nephele_xmpp ? :)
<nephele_xmpp>
update?
<Begasus>
the macos build steps*
<nephele_xmpp>
Not sure that helps
bbby has quit []
<nephele_xmpp>
I was just confused by the order of steps, anyhow
<Begasus>
should be good enough here for Bruno :)
<geniodiabolico>
nephele_xmpp Suppose supertype and subtype are both 64 chars. It gets formatted but the two together plus the forward slash are now 129 chars and so the last 2 get shaved off. I'm not understanding why keeping this destination buffer size the same is more important than actually functioning.
<geniodiabolico>
if it's OK to fail then why not just reject the whole thing when supertype + subtype exceed that size?
<nephele_xmpp>
like waddlesplash said: It would probably be much better to use BString::SetToFormat here, so we avoid using so much stack unnecessarily.
<geniodiabolico>
Does that remove the buffer size stuff?
<geniodiabolico>
like you don't have to predeclare it?
<nephele_xmpp>
If I understand correctly anyhow :D
x10z has joined #haiku
<nephele_xmpp>
for the amend: with the commit you pushes as HEAD commit, just mark stuff you want to add to the commit with “git add” again, and then use git commit —amend and then push it again to the special target (refs for master or how it is called)
<geniodiabolico>
ok, thanks. I will give this a shot.
<geniodiabolico>
I'm trying to generally reduce warnings and a **lot** of them are snprintf and sprintf ones. Is this the preferred general way to do this?
<zard>
I believe the push command is "git push origin HEAD:refs/for/master"
<geniodiabolico>
and is it different answer for Haiku apps vs old legacy library code?
<nephele_xmpp>
I just have a post-it note on my desktop for the git command
dru_satori has quit [Quit: Vision[]: i've been blurred!]
<zard>
If it's in the Haiku repository, it shouldn't be different
<geniodiabolico>
I kind of missed the git era. The last time I delivered production C code we used Subversion. Then I worked at ServiceNow for a long time which didn't use conventional source control. My last day job used git and I fumbled a lot with rebasing and the fiddliness.
<zard>
Note that whatever repositories are meant to be contributed directly from GitHub (e.g. haikuports) will follow a different procedure
<nephele_xmpp>
path max seems to be 1024 though? and not 128?
<zard>
But for the Haiku repository itself, that one follows the steps that you used
<nephele_xmpp>
well, we do use git, but atleast we don’t use the github PR style, so you don’t have to deal with three repositories
<geniodiabolico>
I'll switch to BString and stop thinking about this destination buffer
<geniodiabolico>
I know most of the world loves the PR flow but it sure felt like a lot of overhead to me
FreeFull has joined #haiku
x10z has quit [Ping timeout: 480 seconds]
<nephele_xmpp>
Now i have to figure out how to use dtrace to debug the host tools
<Begasus>
weird, this thing doesn't want to launch if I rename the Deskbar link to LuminanceHDR, a link with luminance-hdr works, both point to the same binary?
<Begasus>
anyway afk for a bit :)
bbjimmy has quit [Quit: Vision[]: i've been blurred!]
diver has quit [Quit: Leaving.]
bbjimmy has joined #haiku
dru_satori has joined #haiku
x10z has joined #haiku
x10z has quit [Ping timeout: 480 seconds]
dru_satori has quit [Quit: Vision[]: i've been blurred!]
<Begasus>
PulkoMandy, navigating in WebPositive doesn't work anymore?
<Begasus>
switching tab made the arrow keys light up to go back, doesn't do the trick now
diver has joined #haiku
<Begasus>
or nephele_xmpp or zard or ... :)
MsInput has quit [Quit: WeeChat 4.5.1]
freddietilley has quit [Quit: WeeChat 4.5.0]
<Begasus>
we're missing a translator for hdr images :)
Anarchos has quit [Quit: Vision[]: i've been blurred!]
mmu_man has quit [Ping timeout: 480 seconds]
HaikuUser has joined #haiku
mmu_man has joined #haiku
diver has quit [Ping timeout: 480 seconds]
<augiedoggie>
i don't think very many Haiku users are running the rsyncd server but we should probably update our package to fix the security problems
<augiedoggie>
those are some pretty serious bugs :P
<Begasus[m]>
time for an update? (never really used it myself, so not that familiar)
<geniodiabolico>
wow
kinkinkijkin has quit [Quit: Leaving]
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
dru_satori has joined #haiku
mmu_man has quit [Ping timeout: 480 seconds]
mmu_man has joined #haiku
<Begasus>
done checking luminance-hdr here I guess, opens HDR images, and I can still click on things :P
<nephele_xmpp>
begasus: back not beeing available was always a bug in the first place
<nephele_xmpp>
I’ll take a look once my gpu stand is back
<Begasus>
well, you would need a library for that to be able to create a translator for it (JXL-Tranlsator is broken for that matter, not sure if it's reported)
<Begasus>
afk
x10z has joined #haiku
<nephele_xmpp>
not sure what you mean… webkit doesn’t use the JXL translator though, it has support directly build in
mmu_man has quit [Remote host closed the connection]
<Begasus>
Tracker can't open them
<Begasus>
same with heif images, but there's no translator for that either :)
<Begasus>
ah, you meant navigating in WebPositive :D yeah, but switching tab worked before
<Begasus>
out now (dogschool)
<Begasus>
cu peeps!
Begasus has quit [Quit: Vision[]: i've been blurred!]
x10z has quit [Ping timeout: 480 seconds]
<Hiryu>
nephele_xmpp: were you ever able to reproduce your build error? No rush at all, I'm just curious if it's a legit issue or a pebkac... But given the binary seems to work otherwise, maybe it's not a pebkac :D
<geniodiabolico>
updated the mime type commits with the BString code
<geniodiabolico>
is there anyhing I need to do in Gerrit or does pushing the commit trigger everything automatically?
<Skipp_OSX>
you need to make sure that you amend the commit and that the Change-Id at the bottom is identical to the one on Gerrit
dodo75 has joined #haiku
dodo75 has quit []
<Skipp_OSX>
The Change-Id is how Gerrit knows to create an updated patch set instead of creating a new patch set.
dodo75 has joined #haiku
<geniodiabolico>
I did check that before I pushed
<Skipp_OSX>
It appears that you have indeed created a PS2 revision on Gerrit, assuming 8837 is you
<geniodiabolico>
yes
<Skipp_OSX>
for future reference, we typically put the component modified at the beginning of the commit, so in this case Mime:
<Skipp_OSX>
if you look at the other commits on Gerrit they say kernel: awk: busses/i2c: etc etc. there's no real standard here but it helps to identify what component is being modified in the log
<Skipp_OSX>
I guess you did that for awk... anyway, a far as the BString formatting goes that looks ok to me.
<geniodiabolico>
ok, so noted
<Skipp_OSX>
for the other awk patch set, it does not seem right to me. You turned gsub( /\"/, "\\\"", classname ) into gsub( /"/, "\\\"", classname ) ... if escaping the " is not needed here, i.e. \" then it's not needed on the right hand side either.
<Skipp_OSX>
hmm maybe not, because the second " is embedded the first one is not, nvm
yann64 has joined #haiku
<dru_satori>
it's always fun to learn new tool chains, but I do so hate inconveniencing established devs when I make mistakes :(
<Skipp_OSX>
no worries
<Skipp_OSX>
The whole point of Gerrit is so that we can get commits into shape before committing them
<Skipp_OSX>
So any mistakes made there are transient so no big deal
x10z has joined #haiku
<dru_satori>
I do get a laugh, I've used GitHub Desktop for so long that I've forgotten how to use the command line altogether :)
MsInput has joined #haiku
x10z has quit [Ping timeout: 480 seconds]
mmu_man has joined #haiku
AlienSoldier has joined #haiku
AlienSoldier has quit []
tqh has joined #haiku
andreaalle has joined #haiku
<andreaalle>
hi all
AlienSoldier has joined #haiku
<andreaalle>
Can I rest easy? Or do I have to prove that I'm okay?
<andreaalle>
you asked me once to do a helloworld ...
<nephele_xmpp>
Hiryu: Yes, I get the same fault
dru_satori has quit [Quit: Vision[]: i've been blurred!]
<Hiryu>
nephele_xmpp: ack, thanks!
HaikuUser has joined #haiku
zard has quit [Ping timeout: 480 seconds]
dru_satori has joined #haiku
jmairboeck has joined #haiku
<Hiryu>
Hmm. Maybe I should try it on FreeBSD and see if the problem exists there... Though there could be a lot of other problems there as well.
<HaikuUser>
hello everyone. Im going to try to do a question about Emacs. I lost my hopes to found someone to use Emacs in Haiku OS but I am one and I have a little problem
<HaikuUser>
Anyone to use Emacs in Haiku?
mikezt has joined #haiku
<mikezt>
hello
<Hiryu>
Why? Is it broken? I run Haiku here in there in a VM, but never enough to try using Emacs (which I actually do use in BSD/Linux)
<HaikuUser>
Hello Hiryu
<HaikuUser>
Emacs works fine, but I can´t use accent words
<HaikuUser>
Im Spanish, the Spanish keyboard works perfectly in Haiku with all programs except Emacs
zardshard has left #haiku [#haiku]
zard has joined #haiku
<dru_satori>
interesting... If I have the NetworkPreferences window open to the wifi devices, it continuously reconnects.
<dru_satori>
I think tonights project will be to move this installation to the nvme from the thumbdrive.
zardshard has joined #haiku
andreaalle has left #haiku [#haiku]
x10z has joined #haiku
zardshard has left #haiku [#haiku]
dru_satori has quit [Quit: Vision[]: i've been blurred!]
zardshard has joined #haiku
zardshard has left #haiku [#haiku]
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
x10z has quit [Ping timeout: 480 seconds]
zardshard has joined #haiku
Anarchos has joined #haiku
zard has quit [Ping timeout: 480 seconds]
mmu_man has quit [Ping timeout: 480 seconds]
diver has joined #haiku
wicknix has joined #haiku
mmu_man has joined #haiku
wicknix_ has quit [Ping timeout: 480 seconds]
DKnoto has quit [Ping timeout: 480 seconds]
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<scanty>
what happened to the channel description & links that used to be in the channel description
<scanty>
on an irc client, obviously
<scanty>
the topic i mean
<scanty>
oh well. bbl
DKnoto has joined #haiku
Anarchos has joined #haiku
<geniodiabolico>
It sure seems like half of all the build warnings come just from unzip
tqh has quit [Quit: Leaving]
<nephele_xmpp>
Hiryu: I doubt it. It worked on intel macos last time i tried
<AlienSoldier>
well, the party is over for google in netsurf :(
<nephele_xmpp>
HaikuUser: the emacs port is pretty much “upstream” you can open an issue at haikuports or email the maintainer about this issue
dru_satori has joined #haiku
nipos has left #haiku [Disconnected: Received SIGTERM]
nipos has joined #haiku
<Skipp_OSX>
well litehtml port coming just in time then
<Skipp_OSX>
I thought the plan was to get Webkit to do HTML email through wrapper classes
dodo75 has quit [Quit: Vision[]: i've been blurred!]
<nephele_xmpp>
That’s still the plan
<nephele_xmpp>
seems silly to ship two html rendering engines when you already have a fully functioning one. But still that is with webkit2, and i doubt that is done *that* soon
<dru_satori>
also, most HTML email rendering engines are less than full featured due to the overhead. Even GMail's html rendering inside of Chrome is limited in what it can actually do due to security issues
<nephele_xmpp>
well, sure we don’t want any remote ressources and such
<dru_satori>
which in turn means you use an engine, but then have to override functionality from the underlying engine.
<dru_satori>
one of the reasons that MS does not use the Edge renderer in Outlook
<phschafft>
over here we filter the HTML to check/eliminate problematic data.
<phschafft>
and pass that filtered HTML to a full engine.
<phschafft>
that also allows adding your standard layout etc.
<phschafft>
(I don't say that is the best way. just that it is another option)
<dru_satori>
yeah. IIRC, Sparkmail does the same
<phschafft>
also there is sandboxing in the HTML standard. which can add another layer of protection.
<nephele_xmpp>
It depends what you want to render I suppose, and what you consider problematic
<nephele_xmpp>
I would not load any networking library at all in the context of the mail app for webkit, only the css+html part of the engine. no js either
<phschafft>
it doesn't do the job for you, but it may still help against the worst if you fail doing the job yourself. kind of safty net.
<phschafft>
nephele_xmpp: I'm with you that e-mail should never load external content. just mind that a lot of 'professional' business include remote loaded images and stuff. so you might want to communicate that to the user.
wicknix has quit [Quit: Leaving...]
<nephele_xmpp>
Yes, I know. I want to do this like iOS email, or improve upon it a bit. Tell you that we didn’t load anything, and offer the option (per email) to load this for “trusted” contacts
<nephele_xmpp>
but then webkit *still* doesn’t get network, it should tell the mail app, and that would fetch a copy and then give a local copy of the image to mail
* phschafft
nods.
<phschafft>
that is about how evolution does it.
<nephele_xmpp>
also, i want to implement pledge so I can mark this webkit as not getting network, ever, so it can be kiled when someone manages to open a socket :D
chaosDNA has joined #haiku
<phschafft>
;)
<Anarchos>
what to do when Debugger complains about missing libstdc++ debuginfo ?
chaosDNA has quit [Remote host closed the connection]
<Skipp_OSX>
skip it
<Anarchos>
Skipp_OSX okay. I try to debug a 'segment violation' on a call to a virtual method, but all seems valid in Debugger :/
<Skipp_OSX>
Can you see if any registers e.g. %eax are set to 0x000000 ?
<Skipp_OSX>
segment violation means something is out of memory range, a lot of the time that means it is NULL which is 0
<Anarchos>
r10 and r10 are equal to '0x00.....01' and r14 and r15 are equal to 0x0
<Skipp_OSX>
that could be normal though, so doesn't help
jmairboeck has quit [Quit: Konversation terminated!]
<Anarchos>
Skipp_OSX and parameters are pointers on the stack, not sure if it is valid or not
<Skipp_OSX>
well clearly one is invalid but it's hard to say which
<Anarchos>
Skipp_OSX it is the code of BootManager, with the bios drive id affected for all drives. So it enters in a loop that was skipped before.
<Anarchos>
Skipp_OSX one parameter has its *_vptr.BPartition without value, the other has is *_vptr.BDiskDeviceisitor with a value 0x908c55cdcc
<Skipp_OSX>
it's possible that _vptr.BDiskDeviceVisitor is used after freed and so that address is bad
<Anarchos>
the BDiskDeviceVisitor is the parent class of a PartitionRecorder which is declared on stack (like that : PartitionRecorder recorder(settings, biosDrive) ) so no new().
<Skipp_OSX>
k so prob not that then
<Skipp_OSX>
address looks valid so it probably is
<Skipp_OSX>
is there anything that is getting new/delete ?
x10z has joined #haiku
<Skipp_OSX>
It has to be something in that loop that is invalid
mmu_man has quit [Ping timeout: 480 seconds]
vdamewood has joined #haiku
<Anarchos>
Skipp_OSX not that i see, but the BDiskDevice is used as a BPartition, though the BDiskDevice path is 'dev/disk/scsi/0/0/0/raw'
<Skipp_OSX>
yeah that seems ok, it's a raw partition, we do that sometimes
mmu_man has joined #haiku
_justin_kelly16668 has joined #haiku
<Anarchos>
Skipp_OSX there is a const_cast<BPartition*> involved too, but i don't know nothing about const_cast
<Skipp_OSX>
const_cast removes const'ness
<Skipp_OSX>
so you can assign a const BPartition* to a non-const one or vice-versa.
<Anarchos>
Skipp_OSX my only guess is that the virtual method _AcceptVisitor is not defined seomewhere.
x10z has quit [Ping timeout: 480 seconds]
_justin_kelly1666 has quit [Ping timeout: 480 seconds]
yuui has joined #haiku
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<dovsienko>
just got a KDL on the latest nightly
Anarchos has joined #haiku
<dovsienko>
VM page fault in AVLTreeBase::_Remove()
<nephele_xmpp>
an assertion?
<dovsienko>
I do not see the beginning of it because the VM is running in a small screen resolution and the backtrace is long enough to wrap over
<dovsienko>
there's RAMFS just before that
<nephele_xmpp>
Okay. Usually this is paged so you should be able to write “bt” again to get thhe start
yuui has quit [Quit: Vision[]: i've been blurred!]
<nephele_xmpp>
but if the system *can* flush to disk it will write this to the syslog, so you can reboot and open a ticket with the snippet from that
<dovsienko>
I was running a Perl script that uses $TMPDIR several times a second
<dovsienko>
I have a screenshot of the entire virtual screen and a serial console open if that helps
<nephele_xmpp>
doesn’t it put the backtrace into the serial console too?
<dovsienko>
I only opened it after the KDL, so the previous messages are not there
<nephele_xmpp>
In any case, since this is in the VM subsystem, and waddlesplash is working on that currently, just open a ticket with what you have now
<nephele_xmpp>
I’d assume you triggered a new assertion
<dovsienko>
but I have the kernel debugger prompt now so can extract diagnostics if required, or even provide remote access to it
<dovsienko>
waddlesplash: would you be interested to debug? should I open a bug report?
<waddlesplash>
run "message" and "bt" in the serial console
<waddlesplash>
to get the original message and backtrace
<dovsienko>
bt is 37 lines long
<waddlesplash>
let's start with a pastebin of it
<dovsienko>
PANIC: vm_page_fault: unhandled page fault in kernel space at 0x0, ip 0xffffffff80129da2
<dovsienko>
waddlesplash: I do not know, this is the first time it occurred, I was just testing a WIP Perl script on various hosts to confirm it is portable
<waddlesplash>
ok
<waddlesplash>
well, I'd reboot and see if it reproduces
<waddlesplash>
you should open a ticket either way but if it's easily reproducible I would like that script :)
<dovsienko>
rebooting
<dovsienko>
KDL about 5 seconds
<waddlesplash>
OK, great
<waddlesplash>
please do open that ticket and I'll investigae
<waddlesplash>
gate
<dovsienko>
this time with AVLTreeBase::Insert(), as in the screenshot
<dovsienko>
open new or update existing?
<waddlesplash>
ah
<waddlesplash>
but is it in the attribute path?
<waddlesplash>
or still SizeIndex
<dovsienko>
updated the file with the backtrace
<dovsienko>
this time I have a complete log from the boot tim
<dovsienko>
(the debugger connection stayed open)
<waddlesplash>
I'd open a new ticket
<waddlesplash>
if it's in fact related then I'll just merge after fixing
<dovsienko>
okay
<dovsienko>
thank you for looking
<waddlesplash>
no problem
x10z has joined #haiku
HaikuUser has joined #haiku
HaikuUser has quit []
<geniodiabolico>
BString.SetToFormat() called on the same variable will reinitialize each time, right? If used multiple times there is no need to clear between uses?
<geniodiabolico>
will it be leaking memory in this situation?
<dru_satori>
hmm. what next. mt792x wlan driver, or fix the Network failing to autoconnect to the last known wifi network when multiple routers exist on the same wlan ?
<dru_satori>
decisions...
<dru_satori>
or wander up to the brewery and drink beer
<nephele_xmpp>
wait, autoconnect failing depends on multiple SSID’s beeing visible?
<Anarchos>
nephele_xmpp for me it is the case.
<dru_satori>
multiple of the same SSID, yes
<dru_satori>
for me, I have 2 mesh networks availalbe ( Orbi and TPLink Deco )
<nephele_xmpp>
Interesting. Would be funky if someone added that to the ticket. But in any case, the proper fix for this probably would also implement homing and grouping severall BSSIDs under the same SSID
<dru_satori>
if I down all but the primary device in the mesh, it autoconnects fine
<Anarchos>
dru_satori i have only 2 different SSID visible, and i never have autoconnect on the SSID i chose.
<dru_satori>
Anarchos lookiing at the code, I think that is simply a bug in which 'open' devices forces the loop to exit.
<waddlesplash>
dovsienko: can you share the script that caused the problem too?
<dru_satori>
not sure yet, haven't dug deep enough, but so long as my SSID is the first ( and only ) instance of it, it works fine
<waddlesplash>
I traced this down a bit on the livestream on Twitch earlier this week after someone asked
<waddlesplash>
it's supposed to be set by wpa_supplicant, if the "persist" flag is set
<dovsienko>
waddlesplash: let me put it online...
<waddlesplash>
but I guess something isn't right there because it doesn't actually get set
<waddlesplash>
dovsienko: thanks
<dru_satori>
excellent. I think I'll wade into that after a bike ride to clear my head of 'work' code
mmu_man has quit [Ping timeout: 480 seconds]
mikezt has quit [Quit: ret]
mmu_man has joined #haiku
<dovsienko>
waddlesplash: it just began to work, but the code is too messy, so I sent you an e-mail with a snapshot of what reproduced it
<waddlesplash>
okay
<dovsienko>
this will go into upstream, but not before I've had a good sleep and cleaned the code up properly
<Hiryu>
Under FreeBSD, Haiku fails to build with several of these errors (with the same header): ../src/add-ons/kernel/file_systems/fat/bsd/sys/endian.h:66:30: error: 'uint32' undeclared (first use in this function); did you mean 'uint32_t'?
<Hiryu>
Maybe it's because I'm using gcc 13?
<nephele_xmpp>
Maybe? have you tried using the defautl clang for the host tools?
<nephele_xmpp>
I assume this is the buildtools
<nephele_xmpp>
Hiryu: the last time the build instructions for FreeBSD were updated… was by me! yay!
<nephele_xmpp>
but it seems i was using gcc9 or something at the time
<nephele_xmpp>
there was no gcc13 then
x10z has quit [Quit: My Unrecognized Mac has gone to sleep. ZZZzzz…]
<Hiryu>
The oldest version of gcc in ports 11... Could try that.
<Hiryu>
Interesting, using uint32 seems to work in other parts of the code (including the same file), which is why I suspect it might possibly be weirdness from using a newer version of gcc
<Hiryu>
*interestingly
<phschafft>
...or the code is just wrong...
<Hiryu>
Hmm. It builds for me under Ubuntu with gcc 13 so that's probably not it
<phschafft>
people love to hate compilers. but the reality is that compilers are virtually right all the time. so if it reports a problem I would always consider the code at fault first.
<Hiryu>
I can't think of many times when the compiler has wrong... I remember a compiler bug that was causing a null pointer derefence or something like that for the Linux kernel because the compiler was skipping the if statement just before it... Probably more than a decade ago
<waddlesplash>
it's probably a wrong ifdef in our code somewhere
PetePete has quit [Ping timeout: 480 seconds]
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<phschafft>
Hiryu: which sounds like some wrong attribute set. I know gcc skips NULL tests when you set the non-NULL attribute. However that is clearly given in the docs, including a statement on why.
<phschafft>
(I don't say that this is the best or most useful thing to do, even while I understand why they did it. just saying that this specific case works as per spec. and people skip the relevant sentence in the spec when reading for some reason ;)
<dovsienko>
compilers do have occasional defects, as most other thing built by humans
<Hiryu>
well, if gcc 13.3.0 works under Ubuntu and not FreeBSD... maybe something is missing from my FreeBSD env?
<Hiryu>
currently building gcc-11
<dovsienko>
but the root cause of not finding an integer type is more likely a missing #include for stdint.h or inttypes.h or some such
<phschafft>
dovsienko: never said otherwise. just that they are still way more often correct than people want. it's what people want vs. reality.
<Hiryu>
the code doesn't do what you want, it does what you tell it
<phschafft>
dovsienko: they use a non-standard type. so I guess it might be defined somewhere within the project.
<phschafft>
Hiryu: ;)
<dovsienko>
which could be a hidden problem because some other header previously included the required header indirectly, but due to changes in Haiku or FreeBSD code the problem could become unmasked
* phschafft
nods.
<Hiryu>
Well, hopefully stumbling across build issues under FreeBSD and MacOS (yesterday) is helpful and not simply me taking up the valuable time of devs
<dovsienko>
perhaps you could identify the root cause and produce a bug fix
<Hiryu>
Perhaps... I don't have a lot of free time. But perhaps I'll be able to figure it out
<phschafft>
what is strange is, that you say the same type does and does not work in different parts of the file.
<phschafft>
(if that is true and not just the error message being skipped for being identical)
<nephele_xmpp>
dovsienko: the package tool is broken on macos ARM for some reason
<nephele_xmpp>
that was the macos build issue anyhow
<nephele_xmpp>
but those are the host tools, and those weren’t really tested on ARM before, so some issues are to be expected :)