Latest Bugs & Tasks

Syndicate content Haiku
Trac Timeline
Updated: 27 min 2 sec ago

Ticket #11160 ([Debugger] Source file path not clickable) created

Tue, 2014-08-26 00:50

As of hrev47617, x86.

Debugger shows "Click to locate source file 'src/gallium/state_trackers/hgl/hgl.c'" however clicking the text results in no source selection window.

Categories: Development

Ticket #11159 (Replace binary config file by plain text config file) closed

Mon, 2014-08-25 20:09

Those are generally flattened BMessages, which can be dumped via the 'message' command.

Categories: Development

Ticket #11159 (Replace binary config file by plain text config file) created

Mon, 2014-08-25 19:41

In latest Haiku release (nightly) most of config files are a species of binary format,
(for example ~/config/settings/ScreenSaver_settings).

As haiku is not BeOS, and as it is free and opensource, why not having all config files in plain text?

It will be more readable, parsable for shell script, and for other uses (tweak).

note : For R1 with legacy BeOS support this cannot be an easy task.
But it can be done for R2 (BeOS will then be run in emulation mode, and some translation of files can be done).


Categories: Development

Ticket #11158 ([PATCH] Template Makefile fixes) created

Mon, 2014-08-25 16:35

As the commit message says: reformatted, rewrapped lines, removed unsupported stuff ("APP_VERSION"), fixed spelling errors and misinformation.

Categories: Development

Ticket #11157 (Slovenian translations are not available) created

Mon, 2014-08-25 15:14

Slovenian translations are not available in Haiku despite being on Pootle since June. Please add them.

Categories: Development

Ticket #11156 (graphics driver blues) created

Mon, 2014-08-25 10:51

hrev47729, real hardware

Integrated intel graphics 83G965

Haiku won't boot even in VESA mode. Forcing the resolution to 1024*768*16bpp started Haiku ok (boot screen+icons look ok). On desktop however, the offset of screen was wrong so that the display started from the center part of the screen and continued to other side.

Maybe somebody could fix the offset for the card? All other gfx modes I tested resulted in garbage.

Categories: Development

Ticket #11155 (WebPositive package missing from repositories.) created

Mon, 2014-08-25 06:51

This problem was introduced in hrev47677.

​ (WebPositive is available)

​ (WebPositive is not in the repos anymore!)

Only the x86_gcc2 repo is affected. I think this is because olta disabled the WebPositive package for gcc2 (which is right, it should be used only on the hybrid side) to fix a warning when building haiku.

Categories: Development

Ticket #11154 (Fundraising status is out of date.) created

Sun, 2014-08-24 18:19

Fundraising 2014

Updated: March 30th

This is just one of many reasons Haiku appears to be a dead project on first glance.

This should be easy to fix. Haiku inc. must keep the bank account up to date. A simple script could update the website from this data.

Categories: Development

Ticket #11104 ([PATCH] JSON parser) closed

Sun, 2014-08-24 17:27

Patch applied, and subsequently some fixes. Thanks a lot, it's already working.

Categories: Development

Ticket #11113 (need checksum for nightly build) closed

Sun, 2014-08-24 16:34

SHA-256 checksums added, thanks zooey (and the NSA).

Categories: Development

Ticket #11153 (KDL in usb module (EHCI)) closed

Sun, 2014-08-24 16:14

Closing as invalid, verified it was a problem on my end.

Categories: Development

Ticket #11153 (KDL in usb module (EHCI)) created

Sat, 2014-08-23 19:03

Recently I resumed development on a wireless usb dongle driver. Last time I tried it, a couple of months ago, I was able to load the firmware and reset the dongle.
Now, every time I open the dongle, I get a KDL in the usb bus manager.
I'm probably doing something wrong, but I wonder if the recent changes in the usb stack broke something.
The KDL shows that EHCI::FinishThread() is called with a NULL parameter, which is then dereferenced in EHCI::FinishTransfer().
Any idea ?

My driver code is here: ​

The usb function which causes the KDL is this one:

size_t actualLength;
status_t status = gUSBModule->send_request(fDevice,
RT2870_RESET, 8, 0, 0, NULL, &actualLength);
if (status != B_OK) {
TRACE_ALWAYS(DRIVER_NAME": firmware reset failed\n");
return status;

In the syslog I get this:
My driver logs:
KERN: ralink_usb: open device net/usb_ralink/0 index 0, KERN: RalinkUSB::Open()
KERN: ralink_usb: selected firmware /system/non-packaged/data/firmware/ralink/rt2870.bin
KERN: ralink_usb: loading firmware...
KERN: ralink_usb: firmware reset...

this is the USB module:
KERN: usb error control pipe 82: timeout waiting for queued request to complete
Last message repeated 1 time
KERN: vm_page_fault: kernel thread accessing invalid user memory!

On boot, the only ehci logs are these:
KERN: usb ehci -1: successfully took ownership of the host controller
KERN: sitd entry size 64, itd entry size 128
KERN: usb ehci -1: successfully started the controller
KERN: add_memory_type_range(563, 0xfc404000, 0x1000, 0)
KERN: usb ehci -1: successfully took ownership of the host controller
KERN: sitd entry size 64, itd entry size 128
KERN: usb ehci -1: successfully started the controller
KERN: usb hub 17: port 1: device removed
KERN: usb hub 22: port 0: device removed
KERN: usb hub 32: port 3: new device connected
KERN: usb ehci -1: fullspeed device connected, giving up port ownership
KERN: usb hub 32: port 4: new device connected

Is the "-1" index correct ?

Categories: Development

Ticket #11152 (Drawing issues of ProcessController menu) closed

Sat, 2014-08-23 16:34

Duplicate. Sorry for the noise...

Categories: Development

Ticket #11152 (Drawing issues of ProcessController menu) created

Sat, 2014-08-23 16:25

This is hrev47724.

When items from ProcessController's menu are removed, white areas appear at the top/bottom. Best seen e.g. in the crowded te