The real bussines use for containers is running/deploying the applications in the same context as they were developed (i.e libraries, running servers required and initial configuration) in my server on the source company.
Also have several instances of the same application but running different configurations (as in web servers + database access), with the same template as origin (using only a few mbs with a diff from the template)
Security isnt going to be granted by containers because they will share the kernel behind them (with the hypervisor), and you cannot guarantee that.
You are probably refering to the isolation feature thatās kinda fuzzy with the containers themselves.
Are there any features found in file systems like ZFS and BTRFS that could be added to BFS? BFS already has many of their features implemented, but is there anything that it doesnāt have? Or is it already at feature parity with both (if not more)?
I keep reading āKVMā in the old-old sense: a {Keyboard + Video + Mouse} hardware switch to share with multiple computers. (>.<)
Then I have remind myselfā¦
Actually same problem with Android Debug Bridge; I keep reading āADBā as Apple Desktop Bus. And I can tell you itās a bazillion times more difficult now to find info on the old Apple tech now, or to find devices that use it. (-^_^-);
Anyway Iāll peruse Git for the login source code and try to implement a āsave sessionā function in the prompt that writes the state of the machine (screen brightness, desktop applet persistence, open applications and the files on which they were operating, and the power state of the machine, etc.) to a file thatās checked before the next login. I need to read the BeBook and Haikuās APIās to understand how to do it correctly though so it will take a while.
Would it be possible to have either Haiku itself or Installer check if the network hardware in a system would require downloading firmware? Itād be nice if thereās a checkbox for installing proprietary blobs in Installer, much like how most Linux distro installers (e.g. Calamares, Ubiquity, etc.) do it. It would reduce the number of newbies who come here asking why their network adapters wonāt work.
Virtually none of the network hardware made in the last 10 years that Haiku supports (Atheros, Intel, all ethernet) requires firmware that is not bundled with Haiku already, anyway.
I would like to see Haiku get hardware acceleration with vulcan or some other, preferably vulcan for compatibility with some proprietary Apps like X-Plane. This would encourage X-Plane developers and other games or graphics programs, to port it to Haiku. A lot of graphical hardware will be accelerated with vulcan to make more use of CPU cores and therefore faster transfer of games physics calculations to graphics cards and there better use. CPUās will probably be more multicore and this is what Haiku is good at: multithreading
i vote for full working a intel graphics driver with screen modesettings, something like developer promise that when you have intel gfx than is OK for haiku.
and i have small wish, like new app for windows 10 that show notifications from android phones on W10 desktop. its posible something similar app for haiku? mtp and storage support.
Hey @juliocfcosta, could you go into the terminal and type in ālistusbā with your tablet connected and then put the output that mentions your tablet here? I could see if your tablet is in the more specific Wacom driver which could be simply lacking your device name in its database.
Better video for video output, split/mirroring different displays, consistency between display resolutions between Legacy & EFI
Chrome or modern Firefox. WebPositive is okay, but something with LastPass support would really help and audio in WP is choppy.
Spotify client, though not as important with a more modern browser.
Encrypted filesystem.
Dropbox client (Iāve seen a project claiming to be this, but itās in Python, so I didnāt take it seriously enough to try it, not sure it would work with 2FA).
Full LaTeX suite, though not as important with a more modern browser (admittedly havenāt tried Overleaf with WP yet)
Amazing items:
GPU-accelerated computing (CUDA, OpenCL)
Native IDE which has advanced Python, Java support. IntelliJ, PyCharm, NetBeans donāt work, are missing features, or are resource-intensive (same apps work fine on Linux on the same hardware, seem sluggish in Haiku)
Porting Web+ to webkit2 would probably fix alot of problems.
Also Firefox has GeckoView on Android (and it is very fast compared to Fennec which operates similar to how Web+ works now using webkit1 except obviously itās the gecko engineā¦ I wonder how hard it would be to adapt that to other OSes so Firefox could be embedded in other applications again.
As far as video drivers that probably requires a lot of work, perhaps hiring someone and supplying them with test hardwareā¦ and that is without even getting to any sort of acceleration. Otherwise it is very slow progress just as people have time and interest.
Actually, some of them will. Ad blocking, for example, is not available through the WebKitLegacy APIs. Geolocation and gamepads are painful to add. Even for SSL warnings I had to hack in some extra stuff into WebKitLegacy to make it possible.
WebKit2 have plugin support AFAIK, and it is plenty times explained: porting another WebBrowser is not a small task, and doesnāt magically solves all your problems. Donāt take the web features granted just because the broser supports it on other platforms, the integration is a really long road.
The phases of porting a webbrowser would look like this:
Some weeks to become acquainted with the build system (1 week, 1 month, possibly even more)
Recognize, that you have to port/re-port/update libraries, build-systems, even programming laguanges (hi ruby!)
Get the build system compile the first source file, meet with platform differences deep in the code. Start to fix it.
Stub out all the gui and other complicated platform dependant code.
Get the browser built in headless mode (2-3 months, probably more)
Write platform-specific gui code (months), probably you have to change/add stuff to Haiku too (app_server, etc)
write/adjust the testcases
Now you have a working demo, which can brokenly render pages and itās crashing only if you move your mouse.
Stabilisation. Years. Do not forget about the users, who will tell: it isnāt working, and you should port their favorite browser instead this crap.
After that you can try to start to enable the features like audio/video support, which also needs platform specific code.
In the maintime upstream diverges from your startpoint every day with great speed, you have to rebase your changes on top of mainline frequently, you have to try to upstream your changes (if they even accept patches for Haiku).
Then, some years later it will be in the same state like WebPositive is today. It is mostly work ok, but crashing or locks up sometimes, while you are bombarded with feature requests (NetFlix working with this browser on Windows, so you only need to add -DDONT_DISABLE_NETFLIX, right?! It should be easy! And btw, you shouldānt port this, but my favorite Browser, when will you do that too?)
Instead painfully go trough this, why donāt you guys trying to help to fix WebPositive? I think we want to have a working browser, not 5 differently broken or differently incomplete.
But thatās my 2 cents, feel free to start to work on any browser, we need diversity on WebBrowser front, but calculate with the listed things.
