Development

Ticket #12633 (Kernel fails to load on OptiPlex 960) closed

Latest Bugs & Tasks - Wed, 2016-02-03 00:52
fixed:

resolved in hrev50070. Let me know if anyone knows why :-\

Categories: Development

platform/bios_ia32: Fix regression introduced in hrev48326

Source Activity - Wed, 2016-02-03 00:48
* Reverts f3e381dd * Details in #12633 * For some unknown reason, this change breaks an Intel Core 2 system from booting (Dell Optiplex 960, E8400) * If anyone has any ideas on why, let me know :-) * Reverts f3e381dd * Details in #12633 * For some unknown reason, this change breaks an Intel Core 2 system from booting (Dell Optiplex 960, E8400) * If anyone has any ideas on why, let me know :-)
Categories: Development

Ticket #12635 (An old BeOS app (source files attached) which no longer properly works on ...) created

Latest Bugs & Tasks - Wed, 2016-02-03 00:29

hrev50067.

The last time that I succesfully run this app called Replitext, was about one year ago, on Haiku.

Replitxt is a transparent text desktop replicant which can display various and dynamic text messages using scripting.

eg: print the current date inside the replicant:

~> time=$(date +"%d/%b/%Y - %R %p ")

Get IDs from the replicants:

~> replitxt --list

Insert the current date into Replitxt:

~> replitxt --text $ReplicantID "Current time: ${time} "

and we'll see:


And so on: as I've said is - was - possible send various inputs, via scripting, to the replicant.
I say *was possible* since in the latest nightly builds, when I attempt to run the command

~> replitxt --list

All that I get is the following error:

shelf_messenger not found (Bad port ID)
There are 0 replitxt replicants out of 0 replicants

I also tried to recompile this app from the source code (it correctly compiles, using setarch x86_gcc2), but the above error is still here.

I attach the zip file with the source code and the binary, maybe someone would be able to fix the code?

Categories: Development

Ticket #12634 (The syslog is continuously flooded with the same messages.) created

Latest Bugs & Tasks - Tue, 2016-02-02 22:49

Hi,
Yesterday I was looking at the folder /var/log/ and I noticed that the syslog file is continuously growing: it never stops. I made a short explicative video (attached to this ticket): in this video we can see the output of "tail -f /var/log/syslog" and the syslog file which continuously grows. It quickly reach the size of 512 KB, the file is created again, and then this strange behaviour is incessantly repeated.
I also attach the syslog itself.

I choose "General" as Component, because I have no idea about the culprit.

hrev50067

Categories: Development

Ticket #12633 (Kernel fails to load on OptiPlex 960) created

Latest Bugs & Tasks - Tue, 2016-02-02 22:46

The kernel fails to load on the Dell OptiPlex 960.

Serial connected shows:

size: 37.6 cm x 30.1 cm
border: 0 cm x 0 cm
Serial Number: G434H87FGB2U
Monitor Name: DELL 1908FP
Horizontal frequency range = 30..81 kHz
Vertical frequency range = 56..76 Hz
Maximum pixel clock = 140 MHz
crtc: h 1328/1440/1688, v 1025/1028/1066, pixel clock 108000000, refresh 6064
Welcome to the Haiku boot loader!
number of drives: 2
add_partitions_for(0x00105420, mountFS = no)
add_partitions_for(fd = 0, mountFS = no)
0x00105648 Partition::Partition
0x00105648 Partition::Scan()
check for partitioning_system: GUID Partition Map
check for partitioning_system: Intel Partition Map
  priority: 810
check for partitioning_system: Intel Extended Partition
0x001057c0 Partition::Partition
0x00105648 Partition::AddChild 0x001057c0
0x001057c0 Partition::SetParent 0x00105648
new child partition!
0x00105648 Partition::Scan(): scan child 0x001057c0 (start = 4194304, size = 629145600, parent = 0x00105648)!
0x001057c0 Partition::Scan()
check for partitioning_system: GUID Partition Map
check for partitioning_system: Intel Partition Map
check for partitioning_system: Intel Extended Partition
0x00105648 Partition::~Partition
0x001057c0 Partition::SetParent 0x00000000
boot partition offset: 4194304
0x001057c0 Partition::_Mount check for file_system: BFS Filesystem
PackageVolumeInfo::SetTo()
PackageVolumeInfo::_InitState(): failed to parse activated-packages: No such file or directory
load kernel kernel_x86_64...
<LOCKUP>

A random mixture of BIOS / boot options tried.

Intel Core 2 vPro

Boot was attempted with intel_extreme driver blacklisted for good measure, no changes.

32-bit and 64-bit kernels both fail with same issue

Categories: Development

Ticket #12632 (My /home/.profile file (with Terminal's customizations) is ignored.) created

Latest Bugs & Tasks - Tue, 2016-02-02 22:28

Hi,
I recently updated Haiku to hrev50067 (previously I had an old nightly image from one year ago; I don't remember the exact hrev, however)

I always had a .profile file under /home, which contains some Terminal customizations (prompt and colors). Well, after the recent update, the /home/.profile file is being ignored: when I open a Terminal window, in fact, my customizations aren't displayed. I can affirm that the content of such .profile is valid, since if I insert the content in the other profile file (the other, the one which is located as /boot/system/settings/etc/profile) I can see my customization.

The following is the content of my .profile file:

Yellow='\e[0;33m\]'        # Yellow
Red='\e[0;31m\]'          # Red
Gray='\e[0;37m\]'         #Gray
export PS1="\[$(tput setaf 7)\][\A]\[$(tput setaf 1)\] « \[$(tput setaf 3)\]baron\[$(tput setaf 7)\]@\[$(tput setaf 3)\]haiku\[$(tput setaf 1)\] » \[$(tput sgr0)\]"
Categories: Development

BMediaEventLooper: Use enqueue_time in a different shape

Source Activity - Tue, 2016-02-02 21:57
* This is the only solution that allowed to use the best of both ways to do this calculus. I've also tested it with a modified sound player that snoozed every time the buffer should be handled, and found that neither of the lateness calculus I tested (including enqueue_time) really solve all problems. That's why I've tried to find an average solution. There's still room for improvements eventually. * This is the only solution that allowed to use the best of both ways to do this calculus. I've also tested it with a modified sound player that snoozed every time the buffer should be handled, and found that neither of the lateness calculus I tested (including enqueue_time) really solve all problems. That's why I've tried to find an average solution. There's still room for improvements eventually.
Categories: Development

Ticket #12631 (UDP serial debug interface) created

Latest Bugs & Tasks - Tue, 2016-02-02 21:36

This ticket logs some ideas PulkoMandy and landonf had in irc about solutions to the "no serial on a lot of new devices" issue.

Debugging Haiku on new hardware can be troublesome, especially if no serial interface is available (common on most modern hardware)

Other operating systems like OS X and windows offer serial debugging via UDP packets. Debugging over UDP would greatly assist users in troubleshooting KDL's, boot issues, or Debugger on system kits.

15:22 < landonf> Ethernet debugging for FreeBSD is actually on my TODO list, so if that coincides with FreeBSD compat driver support on the Haiku side ...                                                          |
15:22 <@kallisti5> landonf: any specs out there for serial over ethernet?                                                                                                                                           |
15:22 <@kallisti5> something next to tcp                                                                                                                                                                            |
15:23 <@PulkoMandy> just use UDP frames, broadcast on some random port                                                                                                                                              |
15:23 < landonf> Most implementations just send ethernet frames                                                                                                                                                     |
15:23 <@PulkoMandy> or I've seen it done with ARP or ICMP headers or whatever                                                                                                                                       |
15:23 <@PulkoMandy> just raw ethernet works as well with appropriate software on the other side of the link anyway                                                                                                  |
15:23 <@kallisti5> http://stackoverflow.com/questions/25317140/serial-data-over-udp-sockets-in-python                                                                                                               |
15:23 < bslsk05> ​stackoverflow.com: Serial data over UDP Sockets in Python - Stack Overflow                                                                                                                         |
15:24 <@PulkoMandy> so… python in the kernel? :D                                                                                                                                                                    |
15:24 <@kallisti5> http://www.andrewhazelden.com/blog/2012/03/setting-up-a-udp-based-network-serial-port-on-a-linksys-wrt54g-router/#Step_2_Copy_the_files_to_your_Linksys_router                                   |
15:24 < bslsk05> ​www.andrewhazelden.com: Setting up a UDP based network serial port on a Linksys WRT54GS router. | Andrew Hazelden's Blog                                                                           |
15:27 < landonf> http://xref.plausible.coop/source/xref/macosx-10.10.1/xnu-2782.1.97/osfmk/kdp/                                                                                                                     |
15:27 < landonf> This is the kernel-side implementation of the protocol OS X uses                                                                                                                                   |
15:28 <@kallisti5> I think just getting kdl into that would be a huge jump                                                                                                                                          |
15:28 <@kallisti5> you can extract syslogs from there                                                                                                                                                               |
15:28 <@PulkoMandy> well, if you get the basic frame send/receive working you could even plug our gdb stub to it                                                                                                    |
15:30 <@kallisti5> what about ethernet drivers after kdl?                                                                                                                                                           |
15:30 <@kallisti5> don't they lock up?                                                                                                                                                                              |
15:30 < landonf> kdp_udp.c provides the UDP implementation of the Kernel Debugging Protocol (KDP). The network driver is polled for packets, on top of which is implemented a *very* minimal ARP/IP/UDP stack.      |
15:31 < landonf> This is what the driver interface looks like: http://xref.plausible.coop/source/xref/macosx-10.10.1/IONetworkingFamily-101/IONetworkController.h#1337                                              |
15:31 < landonf> kallisti5: You have to provide direct/pollled access to read/write of ethernet frames                                                                                                              |
15:32 < landonf> So ethernet drivers have to be extended.                                                                                                                                                           |
15:32 < landonf> And it's not a great solution for debugging the ethernet driver you're using for the debugging interface :)
Categories: Development

[haiku] Re: The state of Haiku (Ingo Weinhold)

Haiku Mailing List - Tue, 2016-02-02 18:45
On10.01.201608:04,AdrienDestugueswrote: OnSun,Jan10,2016at12:19:30AM+0100,MarkHellegerswrote: Hmm,wantedtohavealookatafewWebpositiveissuesfirst. Patchesalsowelcomeinthatarea,ofcourse;) ...
Categories: Development

Ticket #12630 (Late notices are sent continuously even when no audio playing) created

Latest Bugs & Tasks - Tue, 2016-02-02 05:57
  1. Enable tracing in MediaPlayer's AudioProducer.
  2. Run MediaPlayer and start then stop playing something.
  3. "LateNoticeRecieved" prints forever
Categories: Development

Ticket #12629 (MediaPlayer audio output refuses to restart after disconnect) created

Latest Bugs & Tasks - Tue, 2016-02-02 05:53

After disconnecting MediaPlayer's output in Cortex and reconnecting it to the input of an Equalizer, MediaPlayer refuses to play anything again.

Categories: Development

Ticket #12628 ([regression] BBox title text has white edges) created

Latest Bugs & Tasks - Tue, 2016-02-02 03:22

Essentially the same problem as the Deskbar Clock bug.

Categories: Development

BMediaEventLooper: Better solution for negative timestamp

Source Activity - Tue, 2016-02-02 00:10
* This problem happens only at the startup, so that it's preferred that we deal with it in the BMediaEventLooper side. * This solve some audio stability problems I had with slow systems. * This problem happens only at the startup, so that it's preferred that we deal with it in the BMediaEventLooper side. * This solve some audio stability problems I had with slow systems.
Categories: Development

Ticket #12627 (Webpositive crash after while playing html5 videos) created

Latest Bugs & Tasks - Mon, 2016-02-01 22:21

After playing a youtube videos for a while Webpositive crashes. In the debug it has:HTML5 Audio control state: Call (resampling failed). I was testing in hrev50069 x86_64.

Categories: Development

Ticket #12626 (Webpositive crash after while playing html5 videos) created

Latest Bugs & Tasks - Mon, 2016-02-01 22:18

After playing a youtube videos for a while Webpositive crashes. In the debug it has:HTML5 Audio control state: Call (resampling failed). I was testing in hrev50069 x86_64.

Categories: Development
Syndicate content