<Kernspin>
Good morning. I hope a new Beta/Release candidate will be out in the near future. I was trying Beta3 and it left a very good impression to me, in what is achived alreay.
<Kernspin>
I was able to run my Java Swing applications, which I code, by just removing one MacOS routine on Haiku OS with OpenJDK 14. Very nice!
<Kernspin>
applecuckoo: Thanks for the link. So Beta4 is on the horizont. Great.
<applecuckoo>
Yep!
freddietilley has quit [Quit: WeeChat 3.5]
MajorBiscuit has joined #haiku
DKnoto has quit [Ping timeout: 480 seconds]
DKnoto has joined #haiku
nephele has joined #haiku
rennj has quit [Ping timeout: 480 seconds]
freddietilley has joined #haiku
rennj has joined #haiku
applecuckoo has quit [Quit: Time to move on with life. Running Quassel on a Mac.]
tuaris has joined #haiku
tuaris has quit []
freddietilley has quit [Quit: WeeChat 3.5]
freddietilley has joined #haiku
Anarchos has joined #haiku
rennj has quit [Ping timeout: 480 seconds]
Anarchos has quit [Quit: Vision[]: i've been blurred!]
rennj has joined #haiku
<andreasdr[m]>
Yes. This is nice.
Anarchos has joined #haiku
AlaskanEmily has quit [Remote host closed the connection]
rennj has quit [Ping timeout: 480 seconds]
Anarchos has quit [Ping timeout: 480 seconds]
rennj has joined #haiku
<nephele>
I don't suppose there is a magic env var the kernel sets to figure out where a script is located in Haiku?
<nephele>
or maybe a command?
<phschafft>
em?
<nephele>
say, I have a shell script with "#!sh" as interpreter line, and want to work on files relative to the script. I don't have absolute control of where the current working dir is from where the script is called, but i would like it to work on the same files regardless
<phschafft>
that assumes a lot of things.
<nephele>
I could imagine whatever interprets the shebang line could also add a var to the enviroment to indicate the absolute path to the shell script
<phschafft>
like that the filesystem and the view of the filesystem has not changed.
<nephele>
Yes, it does assume quite a bit, but that is fine for my script, it is only ment for a very narrow purpose
<nephele>
(fixing the mime type of lua files in one specific git repo)
<phschafft>
it might be fine for your script, but it is not fine in general. hence such operations are normally not supported. ;)
<phschafft>
the kernel is not written just for your use case. ;)
<nephele>
that argument makes no sense to me
<nephele>
you could say the same about the shebang line allowing command names on haiku
<phschafft>
the difference here is that a shebang CAN be resolved by the kernel at a given point.
<nephele>
shell script isn't ment to be a perfect robust scripting language that covers every edge case perfectly
<phschafft>
while it is not generally true that you can reverse-resolve it at any later point.
<nephele>
the path to the script can be resolved just fine, how else would the kernel have read the shebang line in the first place?
<nephele>
assuming it even is the kernel handeling shebangs
<phschafft>
yes, it can be resolved by the kernel in that very moment. but that doesn't mean that it isn't completly invalid in the next, or may not be resolveable by you.
<nephele>
i fail to see your point, by that argument it would be disallowed to do any file operations because even if you used test to determine if a directory exists it might not exist as soon as you do the next operation to create a file in that dir
<phschafft>
if you assume that the world stops spinning at this point you can also do the forward path. just access $0 or whatever your shell calls it.
<phschafft>
nephele: yes. and this is a huge range of source for critical race conditions and why the *at() system calls have been invented.
<nephele>
Well, if we follow your argument then we should just delete the shell language and be happy
<nephele>
but that seems like a really moot edge case to me, shell *ISNT* safe, at all, and nothing is gained by pretending it is imo
<phschafft>
(and which(1) can then resolve that $0 for you)
<nephele>
no, it cannot
<nephele>
$0 is "shell specific"
<phschafft>
nephele: the point here is that the kernel is not just for shell scripts.
<nephele>
.... and?
<phschafft>
and starting processes isn't specific to shells.
<nephele>
.... and?
<phschafft>
that means that any code that is starting processes must deal with all the other use cases as well.
<nephele>
I really fail to see your point, i asked if there is a functionality for this in haiku, and your argument has been "No because the kernel doesn't do only shell scripts"
<nephele>
Yeah, i didn't ask about starting processes, i asked, explicitly, about shell scripts... that is the system launching something via a shebang line
<phschafft>
nephele: the mechanism behind $0 isn't specific to the shell.
<phschafft>
just that how it is presented to you in your specific shell langauge might be a bit different.
<nephele>
It is, if you run a script from the shell...
<nephele>
some shells will use the exact name you called the script with
<nephele>
like "../../script"
<nephele>
or just "script"
<phschafft>
yes. and that is exactly what the kernel sees.
<nephele>
but that assumes the pwd doesn't change, which isn't always true
<phschafft>
the value is passed from the kernel for exactly your use case.
<nephele>
No, it might also be something completely different
<phschafft>
not by your definition.
<nephele>
I didn't define anything
<phschafft>
plus, again, that is exactly what the kernel uses. the kernel doesn't have anything else.
<nephele>
The common case here is someone clicking a script from the gui, there is much more information available
<phschafft>
you said that you don't care about the view of the file system to ever change. and as long as that holds true that value is correct.
<nephele>
And i fail to see why the kernel would be the one to have to interpret the shebang line
<nephele>
no, again, shells and other processes can put random stuff into it
<phschafft>
if YOU change the value it is changed, sure. but as long as you don't do that it's the correct value.
<phschafft>
basically you said that you don't care about it breaking when you alter things yourself. and now you complain that the value will not be correct if you alter things yourself.
freddietilley has quit [Read error: No route to host]
<phschafft>
I told you what there is, and what everyone else uses in this case. be happy with it or not.
<nephele>
I know of $0, and it has been terribly unreliable for me
wicknix has joined #haiku
<phschafft>
which is exactly what I said in the beginning.
freddietilley has joined #haiku
<nephele>
This is unreliable precisely because there is no proper mechanism to pass this
<nephele>
i asked if there was, and your response seems to be to try and argue with me for no reason about edge cases
<phschafft>
it is unreliable as the asumptions you make don't hold true in a real world system.
<nephele>
No, that isn't true
<nephele>
I asked specifically if there is a mechanism where the OS passes me the full path to the script
<phschafft>
all you do is to define that some edge cases are not edge cases by your taste and therefore should not be consider as such.
<nephele>
whether that changes in the future is irrelevant to wether the OS passes this or not
<phschafft>
that is the full path of the script as known by the kernel.
<phschafft>
end of story.
<nephele>
so far your argument has been "no, use this even more unreliable way"
<nephele>
That makes no sense, if the kernel didn't know the full path there is *literally* no way it could have read the file and determined the shell interpreter, that is assuming that it even is the kernel doing that which seems really unlikely to me
<nephele>
>all you do is to define that some edge cases are not edge cases by your taste and therefore should not be consider as such.
<nephele>
No, i asked about a mechanism, you proceeded to argue about edge cases
<phschafft>
nephele: whatever. no fun to try to help you. maybe someone else can offer you an answer that is more to your taste. have a good day!
<nephele>
SImply answering "I don't know of such a mechanism in Haiku" would have been way more productive than trying to argue with me and telling me to use even worse ways to try and determint the info i would need
wickn1x has quit [Ping timeout: 480 seconds]
<nephele>
i don't consider not answering my question and trying to convince me i'm wrong for asking as "helping" me
mmu_man has joined #haiku
DKnoto has quit [Ping timeout: 480 seconds]
MajorBiscuit has quit [Ping timeout: 480 seconds]
freddietilley has quit [Quit: WeeChat 3.5]
floof58 is now known as Guest2957
floof58 has joined #haiku
Guest2957 has quit [Ping timeout: 480 seconds]
Gaspare has joined #haiku
Anarchos has joined #haiku
MajorBiscuit has joined #haiku
<x512[m]>
phschafft: nephele Shebang is not handled in kernel, it is handled in libroot.so.
<Anarchos>
x512[m] i thought it was in runtime_loader ?
freddietilley has joined #haiku
<x512[m]>
<runtime_loader>test_executable contains shebang detection code, but final executable start arguments are constituted in libroot.so.
freakazoid332 has joined #haiku
frkzoid has quit [Ping timeout: 480 seconds]
rennj has quit [Ping timeout: 480 seconds]
<Anarchos>
x512[m] did you port gtk ?
<x512[m]>
Yes GTK 3 and GTK 4 with my Wayland implementation over Haiku API.
<Anarchos>
x512[m] i am looking for a gtksourceview3 port...
<x512[m]>
For what application?
<Anarchos>
x512[m] it is neeeded to build the ide for coq
<Anarchos>
x512[m] gtksourceview configure complains about AX_IS_RELEASE. Did you encounter this error ?
<matt3>
appshare ?
<matt3>
xD i apologize ...
<matt3>
let me explain: instead of beshare that we share the soul => appshare that we share the apps?
tuaris has joined #haiku
illwieckz has quit [Read error: Connection reset by peer]
<nephele>
matt3: ?
<matt3>
hi nephele
<matt3>
i was thinking about changing name to beshare ...
<matt3>
what do you think?
<bbjimmy>
T%racker is not showinf thumbnails even though the "Media:Thumbnail and Media:Thumbnail:CreationTime attributes are attatched to the file.
<nephele>
I honestly don't know why you would do that, it seems like it would be confusing :D
<nephele>
bbjimmy: I think showing the thumbnails is also tied to the "generate image thumbnails" setting
bjorkintosh has quit [Remote host closed the connection]
bjorkintosh has joined #haiku
Gaspare has quit [Quit: Gaspare]
<matt3>
sheep ide ?
mmu_man has quit [Ping timeout: 480 seconds]
illwieckz has joined #haiku
<Anarchos>
nephele no since few days, thumbnails are enabled by default
bbjimmy has quit [Quit: Vision[]: i've been blurred!]
<Anarchos>
nephele there is a setting indeed, but it is enabled by default.
matt3 has left #haiku [#haiku]
bjorkintosh has quit [Remote host closed the connection]
bjorkintosh has joined #haiku
rennj has quit [Ping timeout: 480 seconds]
bbjimmy has joined #haiku
<Anarchos>
x512[m] what should be my path to compile gtksourceview3 ?
<x512[m]>
As I understand, coqide is written in OCaml so it need GObject Introspection. GObject Introspection build is currently disabled and rebuilding everything using GIR will be required.
<x512[m]>
About GtkSourceView: I recommend to try Meson.
<x512[m]>
configure is obsolete.
rennj has joined #haiku
<Anarchos>
i tried 'meson compile' but it asks for a folder
matt3 has joined #haiku
<Anarchos>
x512[m] coq is written in OCaml; but i think the lablgtk binding from ocaml to gtk uses a C foreign function interface
<x512[m]>
Anarchos: For Meson you need to create build directory first and then call "meson .." from there.
Major_Biscuit has joined #haiku
mmu_man has joined #haiku
MajorBiscuit has quit [Ping timeout: 480 seconds]
bitigchi has joined #haiku
Major_Biscuit has quit [Ping timeout: 480 seconds]
bjorkintosh has quit [Remote host closed the connection]
bjorkintosh has joined #haiku
<matt3>
bjorkintosh : and using a chromebook 4 haiku ?
Gaspare has joined #haiku
<bbjimmy>
waddlesplash It looks like some change messed up the Media:Thumbnail:CreationTime. I have a file with 8/4/55, 11:56AM weeks before I was born.
<matt3>
bbjimmy : check bios date in bios settings ...
<bbjimmy>
it does not display the thumbnail and tracker never adds the Media:Thumbnail attributes on new images.
rennj has quit [Ping timeout: 480 seconds]
<bbjimmy>
matt3 I'll give it a try, but the buis should never return 1955 as a date. and mu timeclock displays the proper date.
<bbjimmy>
*my
bbjimmy has quit [Quit: Vision[]: i've been blurred!]
mmu_man has quit [Ping timeout: 480 seconds]
<Anarchos>
x512[m] i tried to compile the lablgtk package of ocaml, and i got this error : «gtk| Run-time dependency xkbcommon found: NO (tried pkgconfig and cmake)»
<Anarchos>
x512[m] sorry it was not while compiling lablgtk , but gtksourceview, latest revision on master
<Anarchos>
i guess hte next step is «meson install» ?
bitigchi has quit [Ping timeout: 480 seconds]
bitigchi has joined #haiku
rennj has quit [Ping timeout: 480 seconds]
Anarchos has quit [Ping timeout: 480 seconds]
applecuckoo has joined #haiku
<x512[m]>
ninja install
<x512[m]>
meson is needed only for initial configuration or changing options.
applecuckoo has quit []
TurboX has joined #haiku
Anarchos has joined #haiku
rennj has joined #haiku
DKnoto has joined #haiku
<Anarchos>
x512[m] no luck : while the "meson install" succeedeed in libxkbcommon, i still get "../subprojects/gtk/meson.build:439:0: ERROR: Dependency "xkbcommon" not found, tried pkgconfig and cmake" in gtksourceview
bitigchi has quit [Ping timeout: 480 seconds]
BrunoSpr has joined #haiku
xet7 has joined #haiku
Anarchos has quit [Quit: Vision[]: i've been blurred!]
jmairboeck has joined #haiku
KapiX has joined #haiku
<x512[m]>
Meson currently do not properly handle pkgconfig install paths.
<x512[m]>
You can manually set pkgconfig path for meson.