Latest Bugs & Tasks

Syndicate content Haiku
Trac Timeline
Updated: 15 min 12 sec ago

Ticket #12945 ([Screenshot] Display error messages when saving fails) created

Wed, 2016-09-07 12:19

In the process of fixing #10928, I noticed that the screenshot tool doesn't show any errors when saving the image fails (because of insufficient permissions, or if the destination for some reason is not a directory, or any other storage error).

On error, the window just stayed open, instead of closing. This can be very confusing for the users who might not remember that the normal behavior is to close on save, and thus may think the file saved sucessfully when it didn't.

This patch fixes the problem by adding error alerts to the saving function.

Categories: Development

Ticket #10928 ([Screenshot] Artwork folder is read-only (easy)) closed

Wed, 2016-09-07 05:24

Applied in hrev50530, thanks!
Feel free to continue with the improved error handling and user alerts, or something else as you wish :)

Categories: Development

Ticket #12944 ([Debugger] can't load sources for 'make') closed

Wed, 2016-09-07 03:04

Should be fixed in hrev50529.

Categories: Development

Ticket #12944 ([Debugger] can't load sources for 'make') created

Tue, 2016-09-06 09:11

hrev50481 x86_64

I built grep-2.24 and make-4.1 with porter, installed both locally with sources.
Debugger grep lets me locate the grep.c source file successfully, while Debugger make lets me locate the main.c source file but fails to load it.

Categories: Development

Ticket #12943 (BFS ioctl BFS_IOCTL_UPDATE_BOOT_BLOCK integer overflow leading to code ...) created

Tue, 2016-09-06 02:42

In BFS ioctl BFS_IOCTL_UPDATE_BOOT_BLOCK implementation:

			if (update.offset < offsetof(disk_super_block, pad_to_block)
				|| update.length + update.offset > 512)
				return B_BAD_VALUE;

The check on "update.length + update.offset > 512" can overflow 32-bit int. In which case, it will happily memcpy() to an arbitrary memory address.

Sample code:

#include <stdio.h>
#include <fcntl.h>
#include <errno.h>
struct update_boot_block {
	unsigned offset;
	void *data;
	unsigned length;
int main() {
	int fd, ret;
	char data[0x1000] = {0};
	struct update_boot_block block;
	memset(data, 0x42, sizeof(data));
	block.offset = -0x1000;
	block.length = 0x1000; = &data;
	fd = open("/boot/", O_RDONLY, 0);
	ret = ioctl(fd, 14204, &block, sizeof(block));

This will overwrite some kernel memory with 0x42.

(Is it even a good idea to let anyone update the boot block?)

Categories: Development

Ticket #12942 (Graphical issues with the nightly builds) created

Mon, 2016-09-05 12:53

Whenever I try to run the latest versions of the nightly builds from an USB I got graphical issues.

This happened with the following versions:

50527 and 50528.

The problem is that: I run the OS from the usb. Loading Haiku screen shows but when it ends loading and should display the desktop a glitched screen in blue with senseless lines appear (apparently a ghost of what the desktop should be since it changes when I move the mouse or open windows with ctrl+alt+supr.

Tried with the stable release from November 14th, 2012 and it works correctly displaying the desktop and graphics perfectly.

Hardware information: CPU Intel® Pentium®T4500

Categories: Development

Ticket #12941 ({Patch] vfs.cpp: fix clang build) created

Mon, 2016-09-05 11:17

Add '*' at line 1047 [1], to fix build with clang.
[1] ​

/home/haiku/haiku/haiku/src/system/boot/loader/vfs.cpp:1047:14: error: assigning to 'char *' from incompatible type 'char'
                        lastSlash = '\0';
                                  ^ ~~~~
Categories: Development

Ticket #12940 (Debugger should step over instructions, not lines, in dissasembly mode) created

Sun, 2016-09-04 14:57

As in title, when showing disassembly Debugger still steps over C++ code lines, instead of instructions.
In my opinion it should step through each assembly instruction.
Current workaround is to use Run to cursor, but it is inconvenient.

Categories: Development

Ticket #12939 (Displaying of animated Gifs) closed

Sun, 2016-09-04 08:36
Categories: Development

Ticket #12939 (Displaying of animated Gifs) created

Sat, 2016-09-03 17:50

Web browsers can display animated gifs, but the system itself doesn't. Please make this possible.

The same for png animations? I haven't tested it.

Categories: Development

Ticket #12938 (Tapping Sensitivity setting is being ignored) created

Fri, 2016-09-02 22:23

I have Haiku loaded on a Lenovo L530 laptop and even though I have turned the Tapping Sensitivity setting all the way to Off in the Touchpad settings, pressure on the touchpad is still being translated into mouse clicks. It is still quite sensitive and it makes it very hard to do much of anything without "clicking" on a button, menu, etc. without intending to.

The touchpad is a Synaptics UltraNav. Attached is a copy of my syslog.

Categories: Development

Ticket #12937 ([Layer API] Gradients do not respect layer's opacity) created

Fri, 2016-09-02 21:54

When BGradient is applied to a rectangle it renders fully opaque, even though layer's opacity is set to 100 (with BeginLayer).

Screenshot and test program attached.

Categories: Development

Ticket #12732 (Regression with GMA3150 graphics) reopened

Fri, 2016-09-02 12:15

Regression has appeared again with hrev50525 and hrev50526 (it may have appeared earlier I haven't been able to check), it boots the loading screen fine, but when it loads the desktop it becomes all garbled, although it looks as though the system is still functioning.
It boots into a normal desktop if you select fallback VESA drivers.

Categories: Development

Ticket #12928 (Specific tarball causes KDL) closed

Fri, 2016-09-02 08:05
no change required
Categories: Development

Ticket #12936 (CD/DVD: Change startsequence to ask drives past Desktop is loaded) created

Fri, 2016-09-02 06:49

If i have a DVD into my dvd-drive the system need a very long time to load the system.

Is it possible to change asking the DVD/CD/Blue-Ray drive past the Desktop was loaded?

People do not know why the system hangs up at start up.

Categories: Development

Ticket #12935 (Radeon HD5450 no VGA output (regression)) created

Thu, 2016-09-01 23:06

hrev50523 x86_gcc2
ATI Radeon HD5450 with VGA, DVI, HDMI
DVI is working (1680x1050@59.9Hz)
VGA black screen (on-screen error message about unsupported video mode)
attached Radeon_HD5450_listdev
attached Radeon_HD5450_VGA (grep syslog radeon_hd)

DVI and VGA output are both verified working on Alpha 4.1 partition (same PC)

Categories: Development

Ticket #12934 ([Patch] agg_scanline_u.h: fix wrong base_type) created

Thu, 2016-09-01 12:31

As pointed out at [1] and by gcc trunk, modify base_type for scanline32_u8_am from scanline_u8 to scanline32_u8 at line 461 [2].

NOTE: agg 2.4 seems to be continued developing at [3], so we might update our agg from [3].

[1] ​
[2] ​
[3] ​

In file included from /home/haiku/haiku/haiku/src/libs/icon/IconRenderer.h:18:0,
                 from /home/haiku/haiku/haiku/src/libs/icon/style/Style.h:18,
                 from /home/haiku/haiku/haiku/src/libs/icon/message/MessageImporter.cpp:24:
/home/haiku/haiku/haiku/headers/libs/agg/agg_scanline_u.h: In member function 'void agg::scanline32_u8_am<AlphaMask>::finalize(int)':
/home/haiku/haiku/haiku/headers/libs/agg/agg_scanline_u.h:480:32: error: 'agg::scanline_u8' is not a base of 'agg::scanline32_u8_am<AlphaMask>'
/home/haiku/haiku/haiku/headers/libs/agg/agg_scanline_u.h:483:74: error: 'agg::scanline_u8' is not a base of 'agg::scanline32_u8_am<AlphaMask>'
                 typename scanline_u8::iterator span = this->scanline_u8::begin();
/home/haiku/haiku/haiku/headers/libs/agg/agg_scanline_u.h:484:53: error: 'agg::scanline_u8' is not a base of 'agg::scanline32_u8_am<AlphaMask>'
                 unsigned count = this->scanline_u8::num_spans();
/home/haiku/haiku/haiku/headers/libs/agg/agg_scanline_u.h:488:68: error: 'agg::scanline_u8' is not a base of 'agg::scanline32_u8_am<AlphaMask>'
Categories: Development

Ticket #12933 (Intel Graphics - boot with distortionated video after upgrade to hrev50525) created

Thu, 2016-09-01 03:42

Hello. I updated to hrev50525 (x86_gcc2) from a previous nightly from August 15, 2016 (don't remember an exact build number, sorry).

After update, the video looks distorted (looks like when you set a wrong refresh frequency on an old crt screen). The only way to fix it was boot with fail-safe video (VESA).

On previous version (from August 15), the video worked without issues.

My video hardware is:

device Display controller [3|80|0]
  vendor 8086: Intel Corporation
  device 2a43: Mobile 4 Series Chipset Integrated Graphics Controller
device Display controller (VGA compatible controller, VGA controller) [3|0|0]
  vendor 8086: Intel Corporation
  device 2a42: Mobile 4 Series Chipset Integrated Graphics Controller
Categories: Development

Ticket #12932 (HaikuDepot: What is to uninstall, the user should decide) created

Wed, 2016-08-31 17:31

I think it's good that dependent packages of a software are installed over HaikuDepot too. I think it's good that dependent packages of a software deinstalled with the selected program too . But some times I does not want to uninstall all dependent packages.


I have installed 12 yab program on my computer and want to install an new program dependent on a newer yab version. Ater I install the new version of yab on my system, my older yab apps does not run anymore.So I decide to deinstall the latest yab version using HaikuDepot and all 12 programs dependent on yab deinstalled too.

So I need to install not only yab again, I need to install all 12 programs again.

Here I would like to see that the user is given the freedom to decide whether the dependent programs to Uninstall it or not. Because if I want to install an older version of yab from a source other than HaikuDepot because I get about HaikuDepot only the latest version, I would like to be able to decide the installed yab version without disconnecting all of them dependent yab programs. This saves the user a lot unnecessary work.

## German:

Ich finde es gut das über HaikuDepot Software abhängige Pakete mit installiert werden. Ich finde es auch gut das überflüssige Pakete mit deinstalliert werden sollen wenn man ein Programm deinstalliert. Aber manches mal möchte man vielleicht nicht das alle mit einem zu deinstallierten Paket verbundenen Pakete mit deinstalliert werden.


Ich habe um die 12 yab Programm installiert und installiere ein zusätzliches yab Programm welches eine neuere yab Version benötigt auf meinen Rechner. Dadurch läuft nun das neue Programm aber die alten nicht mehr. Nun möchte ich diesen Fehler korrigieren und deinstalliere das zuletzt installierte yab Programm wieder. Dabei erkennt nun HaikuDepot, dass alle yab Programme von dem yab Paket abhängig sind und deinstalliert alle 12 schon vorher installierten yab Programm mit.

Wenn ich nun yab erneut installiere, muss ich danach alle 12 Programme die ich benötige wieder installieren.

Hier würde ich mir wünschen, dass dem Nutzer die Freiheit gegeben wird, zu entscheiden ob die abhängigen Programme mit deinstalliert werden sollen oder nicht. Denn wenn ich eine ältere yab Version aus einer anderen Quelle als HaikuDepot installieren möchte da ich über HaikuDepot nur die aktuellste Version bekomme, würde ich gerne entscheiden können die installierte yab Version zu entfernen ohne alle davon abhängige yab Programme. Das erspart dem Nutzer eine Menge unnötige Arbeit.

Categories: Development

Ticket #12931 (Spirograph (BeOS app) doesn't draw its background transparent) created

Wed, 2016-08-31 02:46

Draws spirographs similar to a popular toy. Problem as stated by packager:

"But the transparency does not work on Haiku, so you will only see the top layer."

I've included some screenshots

Categories: Development