The old setup (serendipity) would work, but it was seriously taxing to the unit, mysql isn't the nicest application to low-memory machines. Wanted to try the 'nanoblogger' setup, so here it is.

So far... it's alright. I find the interface frustratingly clumsy in the usability department (my apologies to the Author), I just don't see why it needs to be so awkward to do things.

I'll tinker with it, rewrite parts here and there.

Now to get email shifted over... {Sigh}

I don't think so.

It still involves recompiling the source code to be able to turn on the joystick/spacenavigator code. This isn't something most people are going to be able to do.

When I (or anyone) can go to the download section, download the SL Viewer for Linux, install it, log in and then enable the Joystick or SN controller from the Preferences menu... THEN it'll be "fixed".

I played with the code myself some (referenced previously)...

What is up with all these big projects using weird and offshoot type programming extensions? To be able to compile SL would involve me finding and installing 10 or 12 additional libraries (in addition to the 50-60 I already have), most of which I've *never* heard of... And I write code myself, so I consider myself somewhat fluent in "what's out there".

Get it in the downloadable tarball... THEN we'll see if it's fixed.

Even if it were still partially broken, there would then be many more people testing and improving it. This was my initial rant to begin with.. it's disabled on linux builds by default. We didn't even GET A CHANCE to use a joystick, broken or not.

Loni

Of course, this machine here isn't the powerhouse saturn is... so the limits are correspondingly lower. duh!

2GB per process on the 64bit AMD machine with oodles of ram. Makes sense.

Will be a while until I can really test this, since saturn is currently down, stress testing my backup/restore system. The root drive has been in there a while, guess it's only natural that it would begin slowly fading... Essentially, the last 'stress' I did... the huge swap file usage, hard drive pounding quaking terror I seemed to have put it though earlier this evening has caused the root drive to begin to give up. SMART began screaming at me during restart, sda is failing! sda is failing! the sky is falling! run for your life!

So a little while later, the system is restoring the backup onto a partition on another drive, rebuild grub and it should be back up within the hour or so. A pain, but that's what good backups are for. Right? And since I use volume labels, nothing needs to be changed since I formatted the new partition with the label 'root' again. taadaaa! (kinda slick that!)

Of course, I have this machine to play with while saturn's restoring, the webserver and email is handled on another machine... waaay too many computers in this apartment. {Grin}. Keeps me off the streets.

And, I remember seeing a sym-link loop in there somewhere in the /var/lib area in the past, so I did rather expect it to get caught and whirr away for a while on that... probably abort with a memory malloc error or something similar.

Well, it *did* abort... but at exactly 16,777,216 items. Doesn't that number ring a bell? Yeah... exactly 16 Meg (16*1024*1024).

Kinda eerie. I've got 4GB of ram, 6 GB of swap... it ate up nearly 7GB of storage real/virtual before it coughed and died. So of course I had to enable core dumps and see....

<run it again!>

Yup... 16,777,216 entries, then:

terminate called after throwing an instance of 'std::bad_alloc'
what(): std::bad_alloc
Aborted (core dumped)

Strange... does 'vector' have an upper limit on how many it can index (beyond memory/ram limits of course)?

So whip up a little test jig, using 'vector <int> tst' just to see how many it can create before it dies...

This time... 536,870,912 items.... gee,that number's familiar too... That's *exactly* 512 Meg (512*1024*1024). Very Weird.

The test jig just stored 'int's, 4 byte values. The original program (dedupe), stores 'md5file' classes...

Hmmm:

512 Meg items @ 4 bytes = 2 GB of storage....
16 Meg items @ ? bytes = ... 2 GB?? (this is a theory...)

Implies the Class is 128 bytes... Let's go add up things and see...

Well... drat. sizeof(md5file) says it's 88 bytes, I'm 40 bytes shy. One-Third too low. Darn. And just for good measure, I verified that the test jig is using 4 byte int's too. Trying 8 byte double's now... testing the theory again. I predict... 256Meg items. Now to see if my theory holds true.

{jeopardy music plays softly in the background}

HEY! 268,435,456 items... 256*1024*1024... 256Meg! The theory seems to hold true...

So what's using the other 40 bytes? {Sigh} Use the Source Luke! Dang-it! I *WROTE* the source...

Not that my machine has anywhere NEAR 16 Million files to tinker with... but still!. Actually, considering that I can predict what it's going to do... maybe it's a feature now. {shrug} dunno.

Had an epiphany while recovering from the last coredump... Efficient memory usage / allocation is usually done by powers of 2... and since 88 > 64, then the next size up is 128. Which would explain the 128 byte per allocation usage predicted by my theory above.

So yeah, it all works. There's not a 16Meg limit to number of items, there's a 2GB limit on the amount of data stored.

Taa Daa!

(Not that this really explains anything, but at least now I know why it's dying at exact 'power-of-two' boundaries.

Loni

(Edit: I figured out the 2G limit too... see next post)

Thought the rest of you might want to try it and see if it helps your experiences with FF3.

There are several bug reports concerning heavy drive thrashing and slow operation. These all boil down to the way FF3 is handling some data.

To help people identify Phishing sites, forgeries (phorgeries?), and general ickiness, FF3 contacts Google.com for EVERY http request. EVERY REQUEST

It then saves the results into 'urlclassifier3.sqlite' in your profile.

This file grew to more than 50MB at which point I read that you could 'vacuum' it using 'sqlite3' to drop the unused database entries. I did this, it shrank to 33MB, a savings of 17MB... but the thrashing continued.

If you will disable the Google checks, then exit FF3 and delete the 'urlclassifier3.sqlite' file from your profile, your FF3 experience will be much nicer, albeit at the expense of some assistance in Phish Site Phinding. (mmm, Ben & Jerry's Phish Food Ice Cream...)

Turn off these two options: (under preferences)

Security -> "Tell me if the site I'm visiting is a suspected Attack site"
Security -> "Tell me if the site I'm visiting is a suspected forgery"

Close the preferences window, exit Firefox, then delete the 'urlclassifier3.sqlite' file:

rm .mozilla/firefox/*/urlclassifier3.sqlite

The file WILL be recreated upon the next start of Firefox3, but it's only 32KB in size, and isn't expected to grow since the two "Tell me" options are off.

It's made a world of difference in the browsing experience for me.

Hopefully this helps others.

Of course, did the fight with the extensions, had to lose a couple, realized I didn't use them really at all, so wasn't a big loss, found a few others providing nice features too.

AdBlock Plus for sure, Forecastbar for weather, Web Developer, Flashblock and FlashGet are pretty much standard fare on all my systems. Others loaded for specific purposes as needed. Found a nice theme too, KDE4+FF3. Pretty!

Anyways, I did my part, downloading the new .bz2 for each computer, so that's three votes from me. It's almost 9am (CDT), and their download counter shows 6.6million++... so they've beat their 1.6M record handily.

It's like Christmas this week... Firefox 3.0 yesterday (Tuesday, 17 June) and OpenSUSE 11.0 on Thursday, 19 June. Ktorrent is idling, all warmed up, waiting for the signal to begin the heavy pulling of the install DVD/CD's.

I used to use Azureus, then realized I didn't use it but once or twice a month (if that!) and Ktorrent had an update that improved it substantially. So AZ went away. I lean on Usenet pretty heavily anymore anyways. (More on that later)

Yes, I ramble... is that bad?',

decode-devices.pl -> src (Perl) (right-click -> save-as)

The output is essentially a verbose parsing of the information. My machine returns this information:

I: Bus=0010 Vendor=001f Product=0001 Version=0100
N: Name="PC Speaker"
P: Phys=isa0061/input0
S: Sysfs=/class/input/input0
H: Handlers=kbd event0 
B: EV=40001 : 18-SND 0-SYN 
B: SND=6 : 2-TONE 1-BELL 

I: Bus=0003 Vendor=046a Product=0048 Version=0111
N: Name="Das Keyboard"
P: Phys=usb-0000:00:02.0-1/input0
S: Sysfs=/class/input/input1
H: Handlers=kbd event1 
B: EV=120003 : 20-REP 17-LED 1-KEY 0-SYN 
B: KEY=7 ff800000000007ff e0b2ffdf01cfffff fffffffffffffffe
   Supports 131 keys
B: LED=1f : 4-KANA 3-COMPOSE 2-SCROLLLK 1-CAPSLK 0-NUMLK 

I: Bus=0003 Vendor=046d Product=c521 Version=0111
N: Name="Logitech USB Receiver"
P: Phys=usb-0000:00:02.1-4.6/input0
S: Sysfs=/class/input/input3
H: Handlers=mouse0 event3 
B: EV=7 : 2-REL 1-KEY 0-SYN 
B: KEY=ffff0000 0 0 0 0
   Supports 16 keys
B: REL=143 : 8-WHEEL 6-HWHEEL 1-Y 0-X 

I: Bus=0003 Vendor=046d Product=c521 Version=0111
N: Name="Logitech USB Receiver"
P: Phys=usb-0000:00:02.1-4.6/input1
S: Sysfs=/class/input/input4
H: Handlers=kbd event4 
B: EV=f : 3-ABS 2-REL 1-KEY 0-SYN 
B: KEY=7fff002c3027 bf00444000000000 1 f808837c000 667bfad9415fed 8e000000000000 0
   Supports 88 keys
B: REL=40 : 6-HWHEEL 
B: ABS=100000000 : 32-VOLUME 

I: Bus=0019 Vendor=0000 Product=0002 Version=0000
N: Name="Power Button (FF)"
P: Phys=LNXPWRBN/button/input0
S: Sysfs=/class/input/input5
H: Handlers=kbd event5 
B: EV=3 : 1-KEY 0-SYN 
B: KEY=10000000000000 0
   Supports 1 key

I: Bus=0019 Vendor=0000 Product=0001 Version=0000
N: Name="Power Button (CM)"
P: Phys=PNP0C0C/button/input0
S: Sysfs=/class/input/input6
H: Handlers=kbd event6 
B: EV=3 : 1-KEY 0-SYN 
B: KEY=10000000000000 0
   Supports 1 key

I: Bus=0003 Vendor=046d Product=c626 Version=0110
N: Name="3Dconnexion SpaceNavigator"
P: Phys=usb-0000:00:02.1-2.3.1/input0
S: Sysfs=/class/input/input12
H: Handlers=event2 
B: EV=20007 : 17-LED 2-REL 1-KEY 0-SYN 
B: KEY=3 0 0 0 0
   Supports 2 keys
B: REL=3f : 5-RZ 4-RY 3-RX 2-Z 1-Y 0-X 
B: LED=100 : 8-MISC 

I: Bus=0003 Vendor=0f30 Product=010b Version=0110
N: Name="Jess Tech GGE909 PC Recoil Pad"
P: Phys=usb-0000:00:02.1-2.3.4/input0
S: Sysfs=/class/input/input13
H: Handlers=event7 js0 
B: EV=b : 3-ABS 1-KEY 0-SYN 
B: KEY=fff00000000 0 0 0 0
   Supports 12 keys
B: ABS=30027 : 17-HAT_0Y 16-HAT_0X 5-RZ 2-Z 1-Y 0-X 

The last two entries are the SpaceNavigator, and the USB Gamepad I've got plugged in for, uh, testing. yeah.. testing.

And for those of you about to squawk... yes, I DO have a "Das Keyboard", and I love it. Keeps most people from bothering my computer too, and has improved my typing speed up further in the triple digits.

Interesting to note that my Logitech mouse somehow (!) has 88 keys on it. Hey... 88 keys... a PIANO?!?

This is in addition to the 16 keys it reports in the previous entry. Really weird!

Anyways...

If you look at the Keyboard entry, and the SpaceNavigator entry, you can see the LED's they support.

The SN supports the 'MISC' labelled LED, actually bit 8, and the keyboard supports KANA, COMPOSE, SCROLL, CAPS, NUM. Bits 4-0.

The leading numbers for each breakdown indicate which bit they are represented by.

1. Device is sensed as a "Multi-Axis Controller". It is quite probable that a small driver module could attach itself as a handler and pass the SN off to the joydev module for seamless joystick usage.
2. Would it need to be a module? Eventually... yes. For now, it could be a small program launched as a daemon to act as a proxy to and from the SN and the JSx device. I'd like to be able to sense all axis inputs, and ALSO toggle the base LED on and off. Of course, since few joysticks have das blinkenlights, that part could quite possibly be non-standard. blech. What if I researched how to send data to joystick as if it were a force-feedback capable device? that's a standard. pretty much. right?
3. "Multi-Axis Controller"... Posted on the web was a short writeup that noted that the device ID blocks for a joystick and a multi-axis controller differed in only a byte or two, or even just a few bits. Tinkering with the SN and a Joystick (JS) show very similar output data too. But I'd have to catch the SN BEFORE the evdev and input modules did. Probably not happening. Besides, that's working too hard and digging too deep to make this work. Items 1 & 2 above should work well.

I'm leaning towards a small daemon that attaches to the event device, and proxies the data to and from a js device. This would be very unobtrusive and require little to no changing of other parts of the system. (talking 'bout you.. Mr. Module!) Would lend itself to being a nice RPM file, a tar.gz/tar.bz2 and a comfy 'configure/make/make install' type setup. Also allow easy tweaking of config items, since we'll likely need to scale the SN motions somehow too. It returns larger values the further it's twisted, pushed and lifted. (Controlling the blue LED is a 'gotta have' thing though. I really don't need another 'glowie LED' on my desk, I've taken to putting dark fingernail polish on the annoyingly bright ones lately)

I'm a long-time Linux user, off and on since the Kernel version started with a zero, and full-time user since around the Y2K fiasco (no, I didn't shift because of that!)

SL opened up their source code for the viewer recently (YAAY!), and I've recently purchased a shiny SpaceNavigator (SN) to play with. Of course, the SN isn't well supported on Linux, but there are a few decent drivers out there, and I've built my own just for fun and learning too...

So then I tackle the issue of joysticks in general, and the SpaceNavigator specifically not working with the SL Viewer in Linux...

One Download of the aforementioned SL Viewer source code later...

I find this:

In the file: .../linden/indra/newview/llviewerjoystick.h

Line 37:

#define LIB_NDOF (LL_WINDOWS || LL_DARWIN)

For those of you who aren't programmers... this essentially says "Define the flag LIB_NDOF if we're compiling for WINDOWS *OR* MAC (darwin)"

The rest of the joystick code then bases its working (or not) on this flag.

example on lines 39: (same file)

#if LIB_NDOF

and in the file:
.../linden/indra/newview/llviewerjoystick.cpp

Lines 168-236:

void LLViewerJoystick::init(bool autoenable)
{
#if LIB_NDOF
... {snip} ...
#endif
}

Which effectively cuts OUT the entire joystick initialization routine. This is continued throughout the joystick support modules.

So... since we're compiling for Linux, LL_WINDOWS and LL_DARWIN are FALSE, while LL_LINUX is TRUE... thus LIB_NDOF is FALSE... and the Linux users [rant mode] HAVEN'T EVEN GOT A CHANCE TO TRY! [/rant mode].

From this, it is apparent that Linux users will NEVER get a joystick working, unless we resort to mapping keystrokes to joystick positions. That's not a solution, it's a crude patch.

Yes, I CAN (and will!) edit the code and enable joystick support for the Linux SL Viewer. This helps me. But doesn't do much for the other 17 Linux users out there.

Please, at least ENABLE the joystick code for Linux users, so we can try to make things work. You're not even giving us a chance.

[Edit] 2008-05-19 02:13

Ok, after reading several bazillion bug tracker entries on SL's bug tracker system (!)... I find that a potential reason this is 'commented out' is because somehow SDL (a Linux Graphics system) doesn't seem to be able to obtain the motion data from the SN.

Hmmm, I'll try writing some sample code tonight (this morning?!?) or tomorrow and see if I can read the data from SDL calls. I think what might be creating havoc is the possibility that SL is wanting to use the 3Dconnexion supplied drivers, or even the FOSS replacement drivers. I've had *NO* problems reading the data using 'evdev' and 'uinput' type code, and really would like to not use those additional drivers. The SN responds as a HID (Human Interface Device) as per USB standards, just like my two USB gamepads, my USB keyboard and USB mice. So I don't think a special driver would be necessary.

I say this based on the fact that my joysticks/gamepads are sensed properly (and without extra drivers) by some of the SDL games and examples I've found. So following that logic... I believe it can be done, and quite simply. (Of course, 'simply' saying that, I've now cursed myself and will suffer through the many circles of Hell while researching this. {sigh})

[/Edit]