Latest Bugs & Tasks
Tried booting the nightly hrev45760 build on a Thinkpad x31 via a Sony USB 2 stick.
The Haiku boot screen icons light and then I get the attached kernel debugging screen.
Sending "hey blah get Frame of View 0 of Window blah" fails if said view # 0 is a BStringView (by opposed to a BButton, BCheckBox ..etc which work very well).
Digging in the code, I see that the base BView::ResolveSpecifier() is overriden by a BStringView::ResolveSpecifier() which returns NULL without calling the base implementation at all (i.e. it disables BView scripting).
resolved in hrev45762
Was originally introduced in hrev38233. Was a solid change, but i'm guessing the abstraction between kFamilies and *family caused a few things to be missed.
Applied fix in hrev45760. Thanks!
Please next time provide a git-format patch.
StyledEdit's Statistics miscalculates the number of words. To reproduce type in a few words and instead of hitting space between them hit enter.
The fix it is easy on line 1828 of StyledEditWindow.cpp is the method IsSpace(fTextView->Text()[i]) which should be changed to IsWhitespace(fTextView->Text()[i]).
I have compiled my local copy and it now counts words correctly.
Should be fixed in hrev45759.
the route command shows negative route prefixes for IPv6 addresses.
A good example without doing any modifications is the -2147454933 for ::1 in route.
If you add an IPv6 IP manually to an interface, you get the same negative number.
Fixed in hrev45753
When i reboot or shutdown Haiku, Deskbar, most of times, will crash. Unfortunately, since Deskbar seems the last component which is closed (also after input_server) i cannot click on "save report" and also, after a reboot, i don't see any message related to Deskbar inside the syslog. In anyway this is reproducible if i quit Deskbar from a Terminal window using:
hey Deskbar quit
After this command, in the syslog i can read these messages:
KERN: vm_soft_fault: va 0x0 not covered by area in address space KERN: vm_page_fault: vm_soft_fault returned error 'Bad address' on fault at 0x0, ip 0x2161e8b, write 0, user 1, thread 0xa8 KERN: vm_page_fault: thread "Deskbar" (168) in team "Deskbar" (168) tried to read address 0x0, ip 0x2161e8b ("runtime_loader_seg0ro" +0xde8b) KERN: debug_server: Thread 168 entered the debugger: Segment violation KERN: stack trace, current PC 0x2161e8b topological_sort__FP7image_tUlPP7image_tUl + 0x4f: KERN: (0x631b3ac0) 0x2162c30 get_sorted_image_list__FP7image_tPPP7image_tUl + 0xb4 KERN: (0x631b3af0) 0x215ed9c terminate_program + 0x28 KERN: (0x631b3b20) 0xf19640 exit + 0x38 KERN: (0x631b3b50) 0x861ff8 _start + 0x64 KERN: (0x631b3b90) 0x21635ae runtime_loader + 0x132 KERN: debug_server: Killing team 168 (/boot/system/Deskbar) KERN: debug_server: TeamDebugHandler::Init(): Failed to get info for team 168: Operation on invalid team KERN: debug_server: KillTeam(): Error getting info for team 168: Operation on invalid team KERN: debug_server: Killing team 168 ()
In BeOS R5 when a BColorControl had keyboard focus you could move between the red, green, and blue ramps with the up and down arrows, and increase or decrease the color value with the left and right arrows. The ramps also had a blue focus ring drawn around them and the currently selected dot in the ramps was drawn as an open circle. I can imagine that keyboard navigation might have worked in palette mode as well.
Fixed in hrev45745
Thanks for the update!
I copied haiku-nightly-hrev45732-x86gcc2hybrid-anyboot to a USB stick over R1alpha4.1. Alpha 4.1 worked fine (excluding sound and power management) but the nightly image prints "loading system" repeatedly and never manages to show the loading stage.
I'm doing this on an Acer Aspire 521 netbook. I don't don't think I can get any more output.
Alpha 4.1 only manages to produce audio if the system has been rebooted straight from Linux, and possibly the BIOS config display. Perhaps things aren't properly initialized under Haiku? Could the filesystem be having a problem with the image being written over a previous Haiku installation?
I'll attach hardware specs reported by Linux.
Should be fixed since hrev45732.
Should be fixed since hrev45732.
Fixed in hrev45732.
The PowerPC boot process panics needing arch_thread_init_kthread_stack implemented. I played around with a slightly modified version of the x86 one, however I just don't understand this still well enough to commit anything (as mine didn't get very far)
This may become easier if we ever get QEMU PowerPC emulation working. (#6071)
Found/fixed some more bugs and optimized things a bit. Also used Ingo's RegExp classes from Ham to implement shell-style wildcard matching rather than just simple string matches. Merged to master in hrev45743.
If anything else is desired as regards these features, please let me know.
Patch applied in hrev45742.
When Fluidsynth will move to an external package, this patch should be kept in mind, BTW.