Here's how to use logcat:
There are two main ways to do a logcat, within android, and through adb.
Logcat within android can be done one of two ways, through a Logcat app:
Here are two good examples are either: aLogcat or Catlog
I prefer catlog, because in my opinion it has a little bit nicer UI. Both of these programs can dump their logs to a txt file, which is very useful for debugging. Or, you can do it in terminal emulator (same rules as running through adb(see below))
From Moscow Desire:
Moscow Desire said:
Just a little more info when doing logcat.
Remember, Logcat will run till you end the session. And it won't always create the file till you do so. It's possible you may not see the logfile for a minute or so.
Edit: For phones/tablets with internal storage & external sd:
(note that the location naming convention may be different depending on device)
Open your terminal app;
Type: logcat > /sdcard/logcat.txt (this should create it in internal memory on the tablet)
To send to ext sd card: logcat > /mnt/external_sd/logcat.txt
I use a tablet for example Phones may have a different naming convention.
MD
Click to expand...
Click to collapse
On the other hand, using adb to run logcat, in my opinion is much more useful, because you can start using it when android boots (i.e. once the boot animation appears.)
The code for logcat to output to a file is
Code:
adb logcat > name of problem.txt
you can also do
Code:
adb logcat -f name of problem.txt
how I prefer to do it is this way:
Code:
adb logcat -v long > name of problem.txt
with the -v flag & the long argument, it changes output to long style, which means every line of logcat will be on its own line (makes it a little neater, imo)
Note: When outputting to a file, you will see a newline, but nothing printed, this is normal. To stop logcat from writting to a file, you need to press ctrl+c.
Here's where using logcat (via adb makes life really easy)
Lets say you find a problem you're having after looking at a logcat.
For example:
When I was trying to use a different ramdisk, wifi wouldn't work so I got a logcat that's almost 1300 lines long (a lot of stuff happens in the background)
So if you are searching for an error in the logcat file (it's always e/ for error, f/ for fatal. Those are the two main things that will break a system.)
Code:
D/dalvikvm( 871): GC_CONCURRENT freed 472K, 6% free 10224K/10823K, paused 1ms+6ms
V/AmazonAppstore.DiskInspectorServiceImpl( 871): Available blocks: 21981, Block size: 4096, Free: 90034176, Threshold: 5242880, withinThreshold? true
D/AmazonAppstore.UpdateService( 871): Received action: null from intent: Intent { cmp=com.amazon.venezia/com.amazon.mas.client.framework.UpdateService }
W/AmazonAppstore.UpdateService( 871): Confused about why I'm running with this intent action: null from intent: Intent { cmp=com.amazon.venezia/com.amazon.mas.client.framework.UpdateService }
D/dalvikvm( 890): GC_CONCURRENT freed 175K, 4% free 9375K/9671K, paused 2ms+3ms
V/AmazonAppstore.ReferenceCounter( 871): Reference (MASLoggerDB) count has gone to 0. Closing referenced object.
E/WifiStateMachine( 203): Failed to reload STA firmware java.lang.IllegalStateException: Error communicating to native daemon
V/AmazonAppstore.UpdateService( 871): runUpdateCommand doInBackground started.
V/AmazonAppstore.UpdateService( 871): Running UpdateCommand: digitalLocker
V/AmazonAppstore.UpdateCommand( 871): Not updating key: digitalLocker from: 1334228488057
V/AmazonAppstore.UpdateService( 871): Finished UpdateCommand: digitalLocker
V/AmazonAppstore.UpdateService( 871): Running UpdateCommand: serviceConfig
V/AmazonAppstore.MASLoggerDB( 871): performLogMetric: Metric logged: ResponseTimeMetric [fullName=com.amazon.venezia.VeneziaApplication_onCreate, build=release-2.3, date=Wed Apr 11 13:10:55 CDT 2012, count=1, value=1601.0]
V/AmazonAppstore.MASLoggerDB( 871): onBackgroundTaskSucceeded: Metric logged: ResponseTimeMetric [fullName=com.amazon.venezia.VeneziaApplication_onCreate, build=release-2.3, date=Wed Apr 11 13:10:55 CDT 2012, count=1, value=1601.0]
W/CommandListener( 118): Failed to retrieve HW addr for eth0 (No such device)
D/CommandListener( 118): Setting iface cfg
D/NetworkManagementService( 203): rsp
D/NetworkManagementService( 203): flags
E/WifiStateMachine( 203): Unable to change interface settings: java.lang.IllegalStateException: Unable to communicate with native daemon to interface setcfg - com.android.server.NativeDaemonConnectorException: Cmd {interface setcfg eth0 0.0.0.0 0 [down]} failed with code 400 : {Failed to set address (No such device)}
W/PackageParser( 203): Unknown element under : supports-screen at /mnt/asec/com.android.aldiko-1/pkg.apk Binary XML file line #16
D/wpa_supplicant( 930): wpa_supplicant v0.8.x
D/wpa_supplicant( 930): random: Trying to read entropy from /dev/random
D/wpa_supplicant( 930): Initializing interface 'eth0' conf '/data/misc/wifi/wpa_supplicant.conf' driver 'wext' ctrl_interface 'N/A' bridge 'N/A'
D/wpa_supplicant( 930): Configuration file '/data/misc/wifi/wpa_supplicant.conf' -> '/data/misc/wifi/wpa_supplicant.conf'
D/wpa_supplicant( 930): Reading configuration file '/data/misc/wifi/wpa_supplicant.conf'
D/wpa_supplicant( 930): ctrl_interface='eth0'
D/wpa_supplicant( 930): update_config=1
D/wpa_supplicant( 930): Line: 4 - start of a new network block
D/wpa_supplicant( 930): key_mgmt: 0x4
(mind you, that's 29 lines out of 1300ish, just for example)
I then could do the following with logcat:
Code:
adb logcat WifiStateMachine:E *:S -v long > name of problem.txt
and this will only print out any errors associated with WifiStateMachine, and anything which is fatal, which makes it about a million times easier to figure out what's going on!
In WifiStateMachine:E, the :E = to look for Errors, the full list of options is as follows:
V — Verbose (lowest priority)
D — Debug
I — Info (default priority)
W — Warning
E — Error
F — Fatal
S — Silent (highest priority, on which nothing is ever printed)
You can replace the :E with any other letter from above to get more info.
In order to filter out anything other than what you are looking for (in this case, WifiStateMachine) you must put a *:S after your last command (i.e. WifiStateMachine:E ThemeChoose:V ... ... AndroidRuntime:E *:S)
Sources: http://developer.android.com/tools/help/logcat.html
http://developer.android.com/tools/help/adb.html
Update for windows users:
Thank go to FuzzyMeep Two, Here's what he's posted for windows
FuzzyMeep Two said:
I have made a tool to simplify this for people. I will send you the .bat version, to see the validity of the file, and post the .exe for everyone else.
http://logcat-tool.googlecode.com/files/logcatHELPER.exe
I made this in about a half hour, so if you fnd any issues let me know. I believe i worked out all functionality issues, and do plan on expanding functionality in the future.
Click to expand...
Click to collapse
(If you used his tool, here's his post, thank him for his work!)
Very nicely done, now i have something i can refer people to
Perfect tutorial. Well done. ... I'll translate your tutorial into my language to share it other people.
pRo_lama said:
Perfect tutorial. Well done. ... I'll translate your tutorial into my language to share it other people.
Click to expand...
Click to collapse
Thank you! And thanks for taking your time to translate it!
If you need me to clarify anything shoot me a PM, and I'll help! (Same goes for anybody else interested in porting to their language)
Pax
Very nice, but I do have a question (and I started programming for Android recently): What do you do when the system reboots while debugging (in my case musb_hdrc.ko)? I tried catlog, but the logs are empty... Is there another way to do that?
Thank you in advance
Krain said:
Very nice, but I do have a question (and I started programming for Android recently): What do you do when the system reboots while debugging (in my case musb_hdrc.ko)? I tried catlog, but the logs are empty... Is there another way to do that?
Thank you in advance
Click to expand...
Click to collapse
Use adb instead of an in-android method to capture the logs, you can start when the boot animation starts (note: Not when the kernel splash screen appears) and when it reboots, it'll automatically end. (i.e. run the following in command prompt:
Code:
adb logcat *:E > oh_nos_it_crashed.txt
)
If you're not getting to the boot animation (keeps cycling on the Kernel splash screen) then you have kernel issues, and you need a kernel that has integrated last_kmsg (a whole different ball of wax that I'm still learning about) to figure out what's going on. (kmesg's are direct kernel output)
Maybe if someone like AdamOutler (or any other Dev who has a much better handle on them than myself) were to explain them, it'd help
Pax
Hey man regarding your guide .
I was testing a built of cm9 from my htc one v and it wasnt booting so the dev told me to do a logcat for it .
Just wondering what the process and commands would be?
Thanks.
Just a little more info when doing logcat.
Remember, Logcat will run till you end the session. And it won't always create the file till you do so. It's possible you may not see the logfile for a minute or so.
But what if you can't connect ADB to your device? You can do this
(note that the location naming convention may be different depending on device)
Open your terminal app;
Type: logcat > /sdcard/logcat.txt (this should create it in internal memory on the tablet)
To send to ext sd card: logcat > /mnt/external_sd/logcat.txt
I use a tablet for example Phones may have a different naming convention.
Nice definition of the display terminology
MD
tranceph0rmer said:
Hey man regarding your guide .
I was testing a built of cm9 from my htc one v and it wasnt booting so the dev told me to do a logcat for it .
Just wondering what the process and commands would be?
Thanks.
Click to expand...
Click to collapse
What you'll wanna do is run the following,
Code:
adb logcat *:E > crash.txt
And look at the output file specifically for lines that say 0x0deadbaad (if you see that, that means there's a problem with the Java libraries it references above it) otherwise look for other F/ lines, that'll point you in the right direction.
If you're having issues where and is not accessible for to buy having USB debugging turned off because of doing a /data wipe (factory data reset) first install a similar Rom (in your case, a prior, stable CM9, boot into it, turn on usb debugging, reboot into recovery & only write /cache & dalvik cache before installing the offending Rom & getting your logcat (sorry if I rambled, just got done working a double )
Pax
Sent from my R800x using Tapatalk 2
Moscow Desire said:
Just a little more info when doing logcat.
Remember, Logcat will run till you end the session. And it won't always create the file till you do so. It's possible you may not see the logfile for a minute or so.
But what if you can't connect ADB to your device? You can do this
(note that the location naming convention may be different depending on device)
Open your terminal app;
Type: logcat > /sdcard/logcat.txt (this should create it in internal memory on the tablet)
To send to ext sd card: logcat > /mnt/external_sd/logcat.txt
I use a tablet for example Phones may have a different naming convention.
Nice definition of the display terminology
MD
Click to expand...
Click to collapse
Thanks, MD!
The problem is when you can't access terminal (bootloops, fc'ing like it's going out of style, etc.) that adb is really useful. Sometimes to trick the system into working, you can (esp with AOSP based roms) install a same version of android (I.e. 4.0.4) & not wipe /data (after turning on usb debugging) & still have an access (I learned this after spending about 2-3 months crack flashing my own builds & getting tired of titanium restoring all my apps)
Pax
Sent from my R800x using Tapatalk 2
Great guide, thanks! Found it through the portal. I just added to my sig so maybe it will help stop some of those questions.
paxChristos said:
What you'll wanna do is run the following,
Code:
adb logcat *:E > crash.txt
And look at the output file specifically for lines that say 0x0deadbaad (if you see that, that means there's a problem with the Java libraries it references above it) otherwise look for other F/ lines, that'll point you in the right direction.
If you're having issues where and is not accessible for to buy having USB debugging turned off because of doing a /data wipe (factory data reset) first install a similar Rom (in your case, a prior, stable CM9, boot into it, turn on usb debugging, reboot into recovery & only write /cache & dalvik cache before installing the offending Rom & getting your logcat (sorry if I rambled, just got done working a double )
Pax
Sent from my R800x using Tapatalk 2
Click to expand...
Click to collapse
Thanks for the info.
Will do that the next time i get stuck at the boot animation
can anyone tell me how to clear / reset logcat?
Thanks in advance
riteshbendre said:
can anyone tell me how to clear / reset logcat?
Thanks in advance
Click to expand...
Click to collapse
Why would you want to do that?
To my knowledge, the only way logcat is reset is if you reboot your phone.
That's why filtering your results is very helpful, because no matter how long it's been running, you can get the info that you want without extraneous information
Pax
Thanks for the clarifications
Man thank you so much for the awesome guide!
Now I can actually filter out my logcat results instead of searching through the entire log.
Sent using Tapatalk
Helpful Tool
paxChristos said:
Here's how to use logcat.....
Click to expand...
Click to collapse
I have made a tool to simplify this for people. I will send you the .bat version, to see the validity of the file, and post the .exe for everyone else.
FIXED HUGE ISSUE CAUSING EMPTY FILES
DOWNLOAD
PM or post here if you run into any issues. They WILL be addressed.
I have made a tool to simplify Logcat creation for people. The creation of this tool was inspired, and guided, by the thread started HERE by paxChristos.
UPDATE 4.1
Corrected an issue causing empty files.
apparently, if the process id contained a space ( 1234) the file would parse correctly, but if the numbers came up to the parenthesis (12345) it would output an empty file named ACTIVITY(12345) with no extension.
This was a huge issue, and i apologize for not seeing it sooner. It has been corrected, and the logcat tool works better than ever. I just ran through a 20,000 line logcat without any empty files appearing.
a couple more things to help with stability.
Underscores (_) will be replaced by dashes (-)
Brackets ([]) and arrows (<>) will be replaced with paranthesis(())
-------------------------------------------------------------------------------------------------------------
HUGE UPDATE
Added the ability to "SUPER PARSE" Files, which separates files into folders by log level and creates a TXT file for each activity
Super parse DOES NOT Work with LONG format Logs.
SUPER PARSE is in its infancy I do expect errors, please let me know if you run into anything.
OLD STUFF
EDIT V3.6 DONE
Added ability to filter logcats by activity name. (logcat -s "FILTER")
Fixed Log Level Setting
EDIT V3.5 DONE
NOW PROPERLY PARSES "LONG" FORMAT LOGCATS.
Thank you to Senior Member jes0411 for pointing out the issues that lead me to create V3.4 and 3.5. If any of you have an issue please let me know. It could be hours before i get to it or weeks, but i will get to fixing it.
EDIT V3.4 DONE
Changed the way the date variable was handled to hopefully fix issues experienced by users of non-English versions of windows.
EDIT V3.3 DONE
EDIT V 3.2 Finished
Google code won't allow any new uploads, so i uploaded the new update to XDA in a ZIP file.
Fixed ADB Location settings, some users were having trouble if they input the ADB folder with any quotation marks in the path name, it will now automatically remove quotation marks to ensure correct syntax.
Fixed issue with ADB Location setting not exiting to previous menu.
Fixed issue with Log Level setting not exiting to previous menu.
Capitalized some stuff
Added a feature that will automatically create the selected folder if no output folder exists and the user attempts to run a log.
Added a check when selecting an output folder that will ask if you want the folder created if it does not exist
EDIT V3.1 DONE
FIXED OPTIONS 7 & 8, URLS WERE MISSING A "?"
CHANGELOG EDITS
SAVES AND PARSES FILES IN TO DIFFERENT OUPUT FILES BASED ON LOG LEVEL (E , I , D , W , ETC)
LOGCAT TOOL v1.1 by FuzzyMeep TWO
I made this in about a half hour, so if you fnd any issues let me know. I believe i worked out all functionality issues, and do plan on expanding functionality in the future.
reposted HERE
Edit: source here.
Batch Source
EDIT: UPDATED TO V1.1
added the framework to add filtering to your logcat, it is in the exe, but not utilized yet (still bugy). i have also worked out a few minor bugs in this release.
Very nice tut. Will try
thanks a ton for the in depth guide. I was always fairly confused on how to approach this, so now I'll have something to refer to in order to help all of the awesome devs here.
mcmb03 said:
thanks a ton for the in depth guide. I was always fairly confused on how to approach this, so now I'll have something to refer to in order to help all of the awesome devs here.
Click to expand...
Click to collapse
+1
Related
There is this linux installer app in the market (link).
As it needs root, and I haven't done it yet, is there anyone here that have tried it??
Thanks,
Thanks, gonna give it a try later.
What about this?
Oh crap, I completely forgot.. Got mixed up with other stuff
Gonna put a reminder.
EDIT: Done. Screenshots attached, one for the standard terminal, the other from aptitude, before updating packages.
Anyone else tried this? I'm having some issues with getting mysql up and running. If anyone is asking well, I'm trying to transform my phone into a server and host my own site
Just because I CAN! *evil laughter*
It seems we now have a computer much more than a smartphone...
Maybe it's time to root to test this.
If it has aircrack-ng then it will worth the work.
Managed to connect via SSH from PC. Still haven't looked into the MySQL issues though.
I finally unbootlocked and rooted my ChaCha and installed this.
I installed Debian, and I'm ready to start testing things.
I installed stable but manually upgraded to testing. But I couldn't setup a SSH server.
Where did you got from??
What do you have in your sources.list?
Update apt from the Distribution Manager and afterwards:
Code:
apt-get check
apt-get update
apt-get upgrade
apt-get install ssh
I still couldn't get MySQL server 5.1 up and running; I have a strange feeling there's something wrong with the distribution, as I've just installed it on my ubuntu box without any errors.
Can you give it a try?
Code:
apt-get install mysql-server-5.1
It should print out something about package not configured yet at some point.
EDIT: Forgot to add
Code:
apt-get install openssh-server
Start the ssh server with
Code:
/usr/sbin/sshd
with "sshd" only you'll get the following error:
Code:
sshd re-exec requires execution with absolute path
Afaik, the sshd $PATH is defined wrong for the service so that's why running it simply with sshd won't work.
Test the functionality afterwards with:
Code:
ssh localhost
Forgot to add the info you requested for sources.list:
Code:
deb http://ftp.debian.org/debian squeeze main non-free
deb-src http://ftp.debian.org/debian squeeze main non-free
deb http://ftp.debian.org/debian squeeze/squeeze main non-free
deb-src http://ftp.debian.org/debian squeeze/squeeze main non-free
Should be the same though I think. What's yours?
Ok. I'll try that.
I already installed psmisc so I could try the lsdev command.
But as it uses /proc and /dev folders, and none of them are in the chroot, they are useless...
But I will try to somehow link from inside the chroot to the actual folders, to try if they work.
EDIT:
I tried mysql server, and I had an error just in the very final stage of setup.
Is this the same for you??
someone got succes?
tokafondo said:
EDIT:
I tried mysql server, and I had an error just in the very final stage of setup.
Is this the same for you??
Click to expand...
Click to collapse
Yep, it's the same.
I'm thinking that mysql 5.0 is already installed and when trying to upgrade to 5.1 a know issue is occurring, where my.cnf does not contain a skip-bdb command at the end, but it doesn't seem to work either; also tried chmod a+r on the my.cnf file as permissions were not granted properly on my device.
@dr. - success on what?
sorry i meant to boot gnome or lxde on htc chacha
dr.ilijevski said:
sorry i meant to boot gnome or lxde on htc chacha
Click to expand...
Click to collapse
Modifying the init scripts in the boot partition, I think it could be done.
Take a look here to have the information that could lead you to do it:
http://www.androidenea.com/2009/06/android-boot-process-from-power-on.html
http://www.androidenea.com/2009/08/init-process-and-initrc.html
Well, that is really the tip of the iceberg, because you would have then to port the X.org server, searching/porting/writing for the drivers for it to work, and finding a desktop enviroment that would run smoothly and with enough room in such a small screen with such a low resolution.
And not only that... What about the radio? Not the FM radio, but the GSM or CDMA radio that allows you to make voice calls and have mobile internet.
Funny thing is that if you check the market app screenshots, there is a screenshot of a running deb dist, apparently on a device.
Alex C. said:
Funny thing is that if you check the market app screenshots, there is a screenshot of a running deb dist, apparently on a device.
Click to expand...
Click to collapse
I sent the developer a couple of questions and had no answer...
This was a kind of experiment, but I think it won't last that much in my ChaCha, because it has no practical use -- for me.
Hmm, got a few extras:
mysqld (the mysql server's main program)
[ERROR]: Can't create IP socket: Permission denied
[ERROR]: Aborting.
So the ld ain't turning on since it has no "permission" to open up sockets, therefore both the mysql-server and the mysql-server-5.1 won't be able to install, since both of them require the mysqld to start first.
Addendum: missing socket file mysqld.sock in /var/run/mysqld/ seems to be the cause of this problem.
Alex C. said:
Yep, it's the same.
I'm thinking that mysql 5.0 is already installed and when trying to upgrade to 5.1 a know issue is occurring, where my.cnf does not contain a skip-bdb command at the end, but it doesn't seem to work either; also tried chmod a+r on the my.cnf file as permissions were not granted properly on my device.
@dr. - success on what?
Click to expand...
Click to collapse
Hi, reviving this question after all this time, aka message in a bottle out to anyone....
so, the stage: debian wheezy fully upgraded in a chroot jail on a nexus 7. ssh, lxde, vnc, zsh, nodejs, erlang, more stuff installed, no problems.
tried to install mysql-server 5.5, got the exact same error as in the image above. drilled a little bit, after the failed install, i ran this:
[1] % sudo mysqld --console --verbose &
[1] 18920
(ssh)[email protected] : ~
[0] % 130415 15:34:48 [Note] Plugin 'FEDERATED' is disabled.
130415 15:34:48 InnoDB: The InnoDB memory heap is disabled
130415 15:34:48 InnoDB: Mutexes and rw_locks use GCC atomic builtins
130415 15:34:48 InnoDB: Compressed tables use zlib 1.2.7
130415 15:34:48 InnoDB: Using Linux native AIO
130415 15:34:48 InnoDB: Initializing buffer pool, size = 128.0M
130415 15:34:48 InnoDB: Completed initialization of buffer pool
130415 15:34:48 InnoDB: highest supported file format is Barracuda.
130415 15:34:48 InnoDB: Waiting for the background threads to start
130415 15:34:49 InnoDB: 5.5.30 started; log sequence number 1595675
130415 15:34:49 [Note] Server hostname (bind-address): '0.0.0.0'; port: 3306
130415 15:34:49 [Note] - '0.0.0.0' resolves to '0.0.0.0';
130415 15:34:49 [Note] Server socket created on IP: '0.0.0.0'.
130415 15:34:49 [ERROR] mysqld: Can't create/write to file '/home/deimosaffair/mysqlhome/var/run/my
sqld/mysqld.pid' (Errcode: 2)
130415 15:34:49 [ERROR] Can't start server: can't create PID file: No such file or directory
note the two last lines. it seems that there are no permissions somehow to create the PID file. i edited the my.cnf file to point to that dir, tried to change the user from "mysql" to "root", nothing worked.
the dir is valid and has wide open access (i made chmod a+rw on it just to be sure), so the permissions should be ok.
the wall is starting to be bloodied from so much headbumping. any ideas in this?
dr.ilijevski said:
someone got succes?
Click to expand...
Click to collapse
Yes, I sincerely apologize for kicking an old thread but I ran into this issue today and nobody seemed to know the answer.
The fix is:
adduser mysql AID_INET
It adds the mysql user to the AID_INET group. After this MySQLd will start just fine.
This is for CyanogenMod and its derivates. Stock ROMs should be fine.
If you are inrested in editing your MAC address, there's a guide linked below. I don't know if stock ROMs will automcatically correct the MAC. In this case I think you have to edit /efs/imei/.nvmac.info (read more something about /efs before playing with it!)
Click to expand...
Click to collapse
The easiest way to do the calibration is:
Turn wifi ON and run
Code:
tiwlan_plts -n
This is the script called from a oneshot init service in my init.latona.rc if someone is interested. It will calibrate and correct the MAC automatically, even after a data wipe:
Code:
#!/system/bin/sh
if [ ! -f /data/misc/wifi/nvs_map.bin ]; then
insmod /system/lib/modules/tiwlan_drv.ko
/system/bin/tiwlan_loader -i /system/etc/wifi/tiwlan.ini -f /system/etc/wifi/firmware.bin
/system/bin/tiwlan_plts -n
rmmod tiwlan_drv
fi
____________________________________
Older method
Click to expand...
Click to collapse
UPDATE: I found an easy way to get the real MAC.
Download the apk attached or run the following commands:
Code:
su
mount -o remount,rw /system
ln -s /system/lib/modules/tiwlan_drv.ko /system/etc/wifi/tiwlan_drv.ko
#Make sure wifi is off
./mfgloader -l #Take the binary from a stock ROM
./mfgloader -u
rm /system/etc/wifi/tiwlan_drv.ko
mount -o remount,ro /system
After that your MAC address shoud begin with D0:C1:B1 and not 08:00:28.
To see your current MAC: System Settings > About phone > Status (turn wifi on)
Click to expand...
Click to collapse
____________________________________
This is the first post with some more informations about the calibration
Click to expand...
Click to collapse
While I was reading some logs, I noticed these lines
Code:
TIWLAN: 1089.979242: Starting to process NVS...
TIWLAN: 1089.983698: No Nvs, Setting default MAC address
TIWLAN: 1089.989007: pHwInit->uEEPROMCurLen: 1c
TIWLAN: 1089.993462: ERROR: If you are not calibating the device, you will soon get errors !!!
...
TIWLAN: 1094.726827: Station ID : 08-00-28-12-03-58
...
You can see that the device is not calibrated and that the default MAC address is assigned.
Each device should have a unique MAC address, two devices with the same MAC address in the same network will give problems.
________
UPDATE
I made two flashable files:
wlan_calibration.zip will only do the calibration
wlan_calibration_rand.zip will do the calibration and will randomize the last three octets of the address
Click to expand...
Click to collapse
If you still want to do it on your own, here the guide:
_______________
WLAN Calibration
Here the steps you need to follow (taken from: WLAN Calibration):
1) Turn wifi off
2) From ADB/terminal emulator run the following commands [1]:
Code:
$ su
# cd /data/misc/wifi
# insmod /system/lib/modules/tiwlan_drv.ko
# start wlan_loader
# ifconfig tiwlan0 up
# tiwlan_cu -b
/ w p 1 l 2 f 2
/ t b v 21
/ t b t 1 0 0 0 0 0 0 0
/ q
# rmmod tiwlan_drv
This will create this file: /data/misc/wifi/nvs_map.bin. If it's not there, you did something wrong.
After that you should not get the error while turning wifi ON.
________
MAC address fix
Even after the calibration, you'll still have the default MAC address (Station ID : 08-00-28-12-03-58). But no problem, you can 'easly' change it: Editing the MAC Address.
You need to get the newly created /data/misc/wifi/nvs_map.bin and edit it with an hex editor as described in the guide linked above.
I still don't know if the calibration will improve anything, but it surely won't hurt:
For optimal Wi-Fi performance it is mandatory to calibrate the Wi-Fi hardware
Click to expand...
Click to collapse
NOTE: nvs_map.bin is in /data, so if you do a factory reset you have to do this again.
___
[1]
ADB: System settings > Developers options (ON) > Root access > Apps and ADB
Wow...good solution buddy. I think u shouldpost this in development thread. People dont look much in general thread.
this should be added to the cm9/10 instalation .zip file
I did flash the zip file, ran the other commands thru terminal and see the file /data/misc/wifi/nvs_map.bin.
I don't understand the next step to be done.
I cannot open the bin file using hexeditor, I gave it root access.
And, in case i open it what mac address should i use? Is some address hard coded to each phone?
Hetalk said:
I did flash the zip file, ran the other commands thru terminal and see the file /data/misc/wifi/nvs_map.bin.
I don't understand the next step to be done.
I cannot open the bin file using hexeditor, I gave it root access.
And, in case i open it what mac address should i use? Is some address hard coded to each phone?
Click to expand...
Click to collapse
I don't know how MAC are assigned, I simply randomized the last part of the address (I kept the first part: 80:00:28).
Anyway I just uploaded a new zip file to make things even simpler. It will automatically randomize the last four three groups. If you want to use it, you need to delete /data/misc/wifi/nvs_map.bin first. EDIT: I changed the script, now you can do the calibration without removing the file.
Anyway I don't know what's wrong with your editor, I did it from my computer.
I tried using hex editor from play store.
I'll try the new zip tomorrow..
As I wrote in the OP, today I found the way to get the original MAC.
I don't know what is needed to make the commands work while in recovery (services needed etc...), so I made a dummy app (I simply adapted the code of an app I made few days ago to do a similar thing).
No fancy icon (it's the default one), no checks, ugly etc... It simply does the job (my Java knowledge is almost zero)
EDIT:
I almost forgot to say that the real MAC is in /efs/imei/.nvmac.info. It was the first place where I looked, but it's an hidden file an I didn't see at first.
bam....
Dear @loSconosciuto
I following this guide to solve the problem that actually I have in my phone with CM 11. In the Status menu, "non available" appears in my device WIFI mac. I following the guide step by step and I couldn't change it.
Could you help me?
carniman78 said:
Dear @loSconosciuto
I following this guide to solve the problem that actually I have in my phone with CM 11. In the Status menu, "non available" appears in my device WIFI mac. I following the guide step by step and I couldn't change it.
Could you help me?
Click to expand...
Click to collapse
This guide is for the kernel 2.6.35 which uses a completely different driver, it's quite an old thread. The calibration is done in a different way and I don't think that's the problem.
I'm not using CM11, so it's quite hard for me to know exactly what's wrong.
loSconosciuto said:
This guide is for the kernel 2.6.35 which uses a completely different driver, it's quite an old thread. The calibration is done in a different way and I don't think that's the problem.
I'm not using CM11, so it's quite hard for me to know exactly what's wrong.
Click to expand...
Click to collapse
Ok, sir. Thank you for your help.
Hi there
trying to fix Samsung s4 active i9295 wifi problem. at least got the MAC address showing (was 02:lots of zeroes) still unable to start wifi. is there anything else I should search for? tiwlan_drv.ko was not there so I downloaded it sepparately. do I need some more modules to be added? using stock 4 file engineering FW Android 5.0.1
Hey-
I was wondering if I could get some help and feedback on a script I conceived to help 4.2 multi-user tree structure to work with pre 4.2 backups (Device, ROMs, Recoveries, Nandroids, Titaniums). Can one of you guru's (_that, becomingx, etc.) chime in and let me know if this seems logical and a good solution, what syntax errors I have, and partially test the theory and code as I am still on 4.1 and locked.
Assuming you have no stock /*/0 directories this script finds all instances of /*/0 (created from the 4.2 multi-user tree structure). /* is "the parent directory" and /*/0 is the "nested directory". First it backs up the parent directory then copies the nested directory (overwrites) into their parent directory. It also combines the backup and the parent directory in the parent directory at the end. This script assumes the user to be single-nested and after running it allows pre-4.2 backups to find their data and targets for restoration. Maybe its OK to combine before the restoration, or after, or not at all? Please give me your feedback on this and don't beat me up too hard as this is my first script with a variable and an array.
Code:
#!/system/bin/sh
nests=( grep ! /*/0 )
for NEST in "${nests[@]}"
do
mkdir /Removable/MicroSD/$NEST_bak
cp -r /$NEST /Removable/MicroSD/$NEST_bak
cp -rf /$NEST/0 /$NEST
#[Restore Pre-4.2 backups]
cp /Removable/MicroSD/$NEST_bak/*.* /$NEST
#[Backup in 4.2]
rm -rf /Removable/MicroSD/$NEST_bak
done
[Edit] Yup, this script needs some serious help. The first line doesn't even work. But when you type it as a command from the prompt it does what I'm looking to capture. How can I put this output in an array?
Code:
grep ! /*/0
Why after running the code above does my prompt which was "[email protected]: #" now show "1|[email protected]: #"? When I was trying to debug the script I even got a 2 and 3|[email protected]: #" Are these some kind of grep shells?
I've now read your post twice, and I still have no idea what you want to do - neither from the script itself nor from the textual description. The confusion starts with the fact that I have no "0" directories on my MicroSD card.
The first step in software development is always to clearly define the problem that should be solved by the software. As long as I don't understand what you want to do, I can't help you with the script - except maybe a hint: "grep" searches for text inside files, "find" searches for files or directories.
Why after running the code above does my prompt which was "[email protected]: #" now show "1|[email protected]: #"? When I was trying to debug the script I even got a 2 and 3|[email protected]: #" Are these some kind of grep shells?
Click to expand...
Click to collapse
The number is the last exit code from the previous command. It indicates that something went wrong.
_that said:
I've now read your post twice, and I still have no idea what you want to do - neither from the script itself nor from the textual description. The confusion starts with the fact that I have no "0" directories on my MicroSD card.
Click to expand...
Click to collapse
No, but you do have some in `/storage/emulated' and `/mnt/shell/emulated' , which I *think* might be what elfaure is referring to (confirm?).
The first step in software development is always to clearly define the problem that should be solved by the software. As long as I don't understand what you want to do, I can't help you with the script - except maybe a hint: "grep" searches for text inside files, "find" searches for files or directories.
Click to expand...
Click to collapse
True words indeed. Although, just telling someone what you want to do and having them show you how to do it, or having them do it for you, is never as much fun as trying to do it for yourself first.
elfaure said:
Hey-
I was wondering if I could get some help and feedback on a script I conceived to help 4.2 multi-user tree structure to work with pre 4.2 backups (Device, ROMs, Recoveries, Nandroids, Titaniums). Can one of you guru's (_that, becomingx, etc.) chime in and let me know if this seems logical and a good solution, what syntax errors I have, and partially test the theory and code as I am still on 4.1 and locked.
Assuming you have no stock /*/0 directories this script finds all instances of /*/0 (created from the 4.2 multi-user tree structure). /* is "the parent directory" and /*/0 is the "nested directory". First it backs up the parent directory then copies the nested directory (overwrites) into their parent directory. It also combines the backup and the parent directory in the parent directory at the end. This script assumes the user to be single-nested and after running it allows pre-4.2 backups to find their data and targets for restoration. Maybe its OK to combine before the restoration, or after, or not at all? Please give me your feedback on this and don't beat me up too hard as this is my first script with a variable and an array.
Code:
#!/system/bin/sh
nests=( grep ! /*/0 )
for NEST in "${nests[@]}"
do
mkdir /Removable/MicroSD/$NEST_bak
cp -r /$NEST /Removable/MicroSD/$NEST_bak
cp -rf /$NEST/0 /$NEST
#[Restore Pre-4.2 backups]
cp /Removable/MicroSD/$NEST_bak/*.* /$NEST
#[Backup in 4.2]
rm -rf /Removable/MicroSD/$NEST_bak
done
Click to expand...
Click to collapse
This is indeed a bit confusing. First I need to know what you're expecting `(grep ! /*/0)' to evaluate to before I can comment on the rest of the code.
[Edit] Yup, this script needs some serious help. The first line doesn't even work. But when you type it as a command from the prompt it does what I'm looking to capture.
Click to expand...
Click to collapse
To capture the output of a command into a variable, you need a dollar sign in front of the opening parenthesis, i.e.:
Code:
nests=$(grep ! /*/0)
But, as _that mentioned, the `grep' call probably isn't the right one to use here. `find' can surely do what you want, but requires busybox to be installed and usually needs to have a bunch of options supplied to it to make sure it's giving you exactly what you want and expect. A simple `ls' command might be all that's needed.
How can I put this output in an array?
Click to expand...
Click to collapse
I, and most other people based on the code I've seen, don't use arrays much in shell scripts. Usually a value separated list is used instead, where the separator value is a space, tab, or newline. Take the example:
Code:
ROOT_LISTING=$(ls /)
for OBJ in $ROOT_LISTING; do
...
done
When the variable $ROOT_LISTING gets expanded at execution time it winds up looking like:
Code:
for OBJ in Removable acct cache [snipped for brevity]; do
and so the `for' loop will iterate through each value after the `in' keyword that is seperated from the next one by a space, tab, or newline. (If we have files, directories, etc. with spaces (possible), tabs (unlikely), or newlines (also unlikely) in the name, we have to be trickier, but I won't get into that here. )
Hopefully this helps. If there's anything else technical that I can yammer on about for the betterment of your understanding, just holler. :good:
Hi Guys-
Thanks for the reply. What service! I simply call for the two guys I need most and like rubbing a magic lamp my genies appear.
When _that reads it twice and still has no clue what you are trying to do you know you need some serious schooling. That's why I'm here - to listen twice as often as I speak and learn something useful.
So here's what I am trying to do. JB 4.2 implemented a radical depart from previous Android versions to the directory tree structure for multi-users as you know, similar to the way Windows uses user profiles to separate users desktops, accessible installed apps, etc. This new structure created /0 subdirectories for certain directories like /sdcard and /sdcard/0 (/data/media and /data/media/0, /clockworkmod and /clockworkmod/0) and others. Playing with grep I discovered I could find all instances of any directory containing a /0 subdirectory (pair). "Find" seems a better approach than "grep" and works exactly the same with no fancy options required.
Code:
find */0
For these directory pairs I'll call them from the above output, the data that was in /directory in 4.1 is now located and pointed to in /directory/0 in 4.2. I figured if I could find all these pairs, and copy /directory/0 into /directory (after backing up /directory to SD) then 4.1 backups looking for data in /directory could now find their data and targets. Then I could copy the SD backup of /directory into /directory (now overwritten by /directory/0) to merge everything into a single folder. But after thinking about this again, this is incorrect. What I meant to do at this step is copy all /0 subdirectories of a given /directory/0 into that folder [ /directory/0/0 into /directory/0 and directory/0/0/0 into /directory/0 and /directory/0/0/0/0... into /directory/0 ] to combine all these sub-nested directories into a single /0 directory. Deep subnesting causes by crack flashing custom ROMs over and over on JB 4.2. This is the concept and I'm trying to implement a script to do it.
So I will attempt to replace grep with find and the array with a variable value space separated list and see what I come up with. Becomingx, you are right about wanting to use my brain a bit to figure it out on my own, even if it sends me down a one-way road to a dead end. If I learned something in the process then its all good and I can always ask for directions at the locked gate if I arrive there, but even with 4 flat tires and a broken axle I'm still moving (forward hopefully) as long as there's still road to travel on.
Thks
So you want to write a script to downgrade from 4.2 to 4.1?
Test new avitar
How do you delete a post? Under Edit/Delete I only find Edit??
_that said:
So you want to write a script to downgrade from 4.2 to 4.1?
Click to expand...
Click to collapse
No, I want to write a script that lets you restore a pre 4.2 backup (Nandroid, Titanium, Recovery, ROM) on 4.2 without having to move/copy files.
elfaure said:
No, I want to write a script that lets you restore a pre 4.2 backup (Nandroid, Titanium, Recovery, ROM) on 4.2 without having to move/copy files.
Click to expand...
Click to collapse
Are you sure that this is not already handled by TWRP, or by TB?
Recovery and ROM themselves are not affected by the /0 move in /data anyway.
Thanks becomingx, getting somewhere now but I still have problems.
When I type this at the terminal I get what i am looking for as output
Code:
su
cd /
find */0
Output: (These are test directories I created)
data/0
sdcard/0
sdcard/0/0
But when I run this script, the above output is repeated 3 times? Why is that?? One for each file system /(rootfs), /data, and /sdcard? How can I rewrite the script to produce the same output as from the terminal?
Code:
#!/system/bin/sh
cd /
NEST=$(find */0)
for OBJ in $NEST; do
echo "$NEST"
done
_that said:
Are you sure that this is not already handled by TWRP, or by TB?
Recovery and ROM themselves are not affected by the /0 move in /data anyway.
Click to expand...
Click to collapse
Long story short, yes if you had all the latest and greatest installed before you upgraded to 4.2 then its a transparant issue. But if you're already "nested" because you used an older custom recovery and factory reset a few times to crack flash some some 4.2 ROMs or need to recover to a previous Android version then there is a use for this script. If I can learn how to script variables and help some nested folks out at the same time then _that's what I call a win-win. I'm not on 4.2 nor do I plan on going there anytime soon so this won't help me out less what I learn by doing it, although it is interesting and challenging to try to implement an automated solution for. Plus I'm like you guys, I just really enjoy rooted Android and helping people with it when I can (within my very limited but expanding skill set).
By the way, what is the name and /data path to that special file?
****************************************************************************************************
http://androidforums.com/verizon-ga...t/649940-4-2-sdcard-sdcard-0-observation.html
http://androidforums.com/verizon-ga...y-bean-roms-edited-3-24-13-a.html#post5796630
Quoted from Androidspin.com -
"With Android 4.2, Google introduced multiple users as a new feature. In order to accommodate multiple users, Google is now giving each user a their own folder for storage. If you upgraded to 4.2 from 4.1, then the 4.2 ROM will look for a certain file in /data to determine whether it needs to migrate all of your files to the new multi-user data structure. By default, 4.2 migrates all of /data/media to /data/media/0.
A problem arose though with custom recoveries. A custom recovery retains the /data/media folder during a factory reset. When you factory reset and then boot a 4.2 ROM again, the 4.2 ROM will migrate everything in /data/media again. It will migrate your files every time you factory reset. This multiple migration is what resulted in some people having their files moved to /sdcard/0 or even /sdcard/0/0 etc.
In TWRP 2.3.2.0 we have corrected this problem by ensuring that we do not delete the special file during a factory reset. However, if ended up having your files upgraded you will need to move or merge them back into /sdcard. Also, if you have moved your TWRP folder from /data/media/0 to /data/media so that you could restore backups while using prior TWRP versions, you may now need to move the TWRP folder back into /data/media/0.
As a special note, if you restore a backup to a prior version of Android, you may have to move your files out of /data/media/0 and into /data/media to be able to see them again."
elfaure said:
TBut when I run this script, the above output is repeated 3 times? Why is that?? One for each file system /(rootfs), /data, and /sdcard? How can I write the script to produce the same output as from the terminal?
Code:
#!/system/bin/sh
cd /
NEST=$(find */0)
for OBJ in $NEST; do
echo "$NEST"
done
Click to expand...
Click to collapse
Read your script - for each item in $NEST you output all of $NEST instead of the single item $OBJ.
_that said:
Read your script - for each item in $NEST you output all of $NEST instead of the single item $OBJ.
Click to expand...
Click to collapse
Ah ha. Lost sight of the fact that a multi-object variable really behaves like an array and you can call its objects as independent variables (syntax wise; I thought that was what I was doing with echo $NEST). Variables in a variable. That's why I started with an array. Thanks.
Code:
#!/system/bin/sh
cd /
NEST=$(find */0)
for OBJ in $NEST; do
echo "$OBJ"
done
Output:
data/0
sdcard/0
sdcard/0/0
New Questions:
How can I strip off all the "/0's" from each object variable and store it in a new mutil-variable until there are no more "/0's to strip off and consolidate duplicates? The desired result here would be VAR=$(data, sdcard).
If two or more object variables have the same root directory name, how can I select only the object variable with the most /0's between two variables (sdcard/0 and sdcard/0/0; I would want only the latter) then store those object variable in a new variable including all other object variables? The desired result here would be VAR1=$(data/0, sdcard/0/0)
Code:
#!/system/bin/sh
nests=( grep ! /*/0 )
for NEST in "${nests[@]}"
do
mkdir /Removable/MicroSD/$NEST_bak
cp -r /$NEST /Removable/MicroSD/$NEST_bak
cp -rf /$NEST/0 /$NEST
#[Restore Pre-4.2 backups]
cp /Removable/MicroSD/$NEST_bak/*.* /$NEST
#[Backup in 4.2]
rm -rf /Removable/MicroSD/$NEST_bak
done
becomingx said:
To capture the output of a command into a variable, you need a dollar sign in front of the opening parenthesis, i.e.:
Code:
nests=$(grep ! /*/0)
[snipped]
I, and most other people based on the code I've seen, don't use arrays much in shell scripts. Usually a value separated list is used instead...[snipped]
Click to expand...
Click to collapse
"nests" in my original code here refers to an array not a variable. Assuming that, is the syntax correct (for reference)? I like the multi-item(object) variable value separated list you showed me and _that corrected my use of and will use that instead for this script.
elfaure said:
For these directory pairs I'll call them from the above output, the data that was in /directory in 4.1 is now located and pointed to in /directory/0 in 4.2. I figured if I could find all these pairs, and copy /directory/0 into /directory (after backing up /directory to SD) then 4.1 backups looking for data in /directory could now find their data and targets. Then I could copy the SD backup of /directory into /directory (now overwritten by /directory/0) to merge everything into a single folder. But after thinking about this again, this is incorrect. What I meant to do at this step is copy all /0 subdirectories of a given /directory/0 into that folder [ /directory/0/0 into /directory/0 and directory/0/0/0 into /directory/0 and /directory/0/0/0/0... into /directory/0 ] to combine all these sub-nested directories into a single /0 directory. Deep subnesting causes by crack flashing custom ROMs over and over on JB 4.2. This is the concept and I'm trying to implement a script to do it.
Click to expand...
Click to collapse
Wow! This is very messy, no wonder you're trying to automate the cleanup.
So I will attempt to replace grep with find and the array with a variable value space separated list and see what I come up with. Becomingx, you are right about wanting to use my brain a bit to figure it out on my own, even if it sends me down a one-way road to a dead end. If I learned something in the process then its all good and I can always ask for directions at the locked gate if I arrive there, but even with 4 flat tires and a broken axle I'm still moving (forward hopefully) as long as there's still road to travel on.
Click to expand...
Click to collapse
Love the analogy here :laugh:
elfaure said:
By the way, what is the name and /data path to that special file?
Click to expand...
Click to collapse
Looks to be `/data/.layout_version' [ref].
elfaure said:
Code:
#!/system/bin/sh
cd /
NEST=$(find */0)
for OBJ in $NEST; do
echo "$OBJ"
done
Output:
data/0
sdcard/0
sdcard/0/0
New Questions:
How can I strip off all the "/0's" from each object variable and store it in a new mutil-variable until there are no more "/0's to strip off and consolidate duplicates? The desired result here would be VAR=$(data, sdcard).
Click to expand...
Click to collapse
If we're going full steam ahead with the requirement of Busybox being installed, then this is pretty easy:
Code:
VAR="$(find */0 | sed -e 's!/0!!g' | sort -u)"
Sed is our Stream EDitor. The option `-e' tells sed that what follows is a sed script.
The script itself (s!/0!!g) breaks down like this:
* `s' means we are doing a substitution.
* The exclamation point (!) is just a delimeter; any character can be used. Typically a forward slash (/) is used, but since directory paths have forward slashes in them, we go with something else to make it easier to read. Otherwise we'd have to escape the forward slashes and it would look like this: s/\/0//g
* `/0' is what we are looking to match.
* Another exclamation point (!) indicates we are done with our "search" pattern and next begins our "replace" pattern.
* Since we are just looking to just get rid of all occurences of `/0', our replacement pattern is literally nothing.
* Another exclamation point (!) to indicate the end of the replacement pattern.
* Finally, `g' (for global) tells sed to keep applying the substitution until it no longer matches. This means a path like `/dir/0/0/0' gets each `/0' stripped away one at a time until there are none left.
The last command sorts our listing so that the `-u' (for unique) option can be applied to it, which removes duplicate lines.
If two or more object variables have the same root directory name, how can I select only the object variable with the most /0's between two variables (sdcard/0 and sdcard/0/0; I would want only the latter) then store those object variable in a new variable including all other object variables? The desired result here would be VAR1=$(data/0, sdcard/0/0)
Click to expand...
Click to collapse
The easiest way to do this is just check for the presence of a `0' dir at each level. If one exists then there's another level of nesting below us and we ignore the current level:
Code:
VAR1=""
for i in $(find */0); do
if [ ! -d $i/0 ]; then
VAR1="$VAR1 $i"
fi
done
Wow! Thanks!! But now I'm suffering from cognitive dissonance (dis-sed-ance). I knew it would require either find, grep, awk, or sed but using those commands without having any grip on regex is nearly impossible. Can you point me to a good resource for learning Android regex or is any GNU/Linux resource valid? Is this a good one?
http://www.linux.org/article/view/introduction-to-regular-expressions-within-a-shell
elfaure said:
I'm not on 4.2 nor do I plan on going there anytime soon so this won't help me out less what I learn by doing it
Click to expand...
Click to collapse
So does that mean you don't have any "0" directories in your /sdcard right now, and you are doing all this just for fun?
At least, after reading your additional material, I am now beginning to understand the problem.
However I don't understand how you get any useful output from "find */0" - I get either "No such file or directory", or a listing of thousands of files, depending on where I start this.
elfaure said:
Can you point me to a good resource for learning Android regex or is any GNU/Linux resource valid?
Click to expand...
Click to collapse
The nice thing with regular expressions is that the basics are the same everywhere.
But I don't know if you even need such powerful tools... Correct me if I am wrong, but your script should convert this structure:
/data/media/0
/data/media/0/stuff1
/data/media/0/0
/data/media/0/0/stuff2
/data/media/0/0/0
/data/media/0/0/0/stuff3
to
/data/media/0
/data/media/0/stuff1
/data/media/0/stuff2
/data/media/0/stuff3
?
So you'd basically need to run this inside /data/media/0:
1. If there is no subdirectory named "0", there is nothing to do, and we can exit
2. If there is a subdirectory named "0", move everything from inside one level out
2a. how do you want to handle conflicts, if the file or directory you want to move outside already exists in the parent directory?
3. go to step 1
[Reply to _that's last post. Getting lazy, I'm at lunch...]
Here is my output from
Code:
find */0
[email protected]:/ $ su
[email protected]:/ # find */0
data/0
sdcard/0
sdcard/0/0
[email protected]:/ # exit
[email protected]:/ $ find */0
data/0
find: data/0: Permission denied
sdcard/0
sdcard/0/0
1|[email protected]:/ $
I'm not sure about the structure I am looking for until I get some output from this code run on a 4.2 device with nested /0's so I can see all of them and their variations. I also want to see what happens to subsequent user directories nested starting with /data/media/10 (second user) and /data/media/11 (third user) and how they nest in comparison.
The only /0 directories that I have on my device were created by me to test this script. I am still on 4.1 and this is a 4.2 only issue so as you state its "just for fun" or rather "for the fun of learning" but at least its not a "hello world" script that does nothing useful. I'm hoping to automate someone's cleanup after installing 4.2 with this, maybe my own at some point, if required.
Here's some more info about the issue:
http://teamw.in/DataMedia
ps-I just discovered the equivalent to the stock browser "about:debug" for Chrome is "about:flags" which brings up an extensive list of experimental settings. I enabled all that were GPU and speed related and its a marked improvement (it feels 25-75% faster). With a 100MB Chrome cache I can have about 20 tabs open at the same time now and its still fast and snappy.
Not sure if this will work or not, but if you want to try to retain single user mode in 4.2 you can try this:
http://androidforums.com/verizon-ga...y-bean-roms-edited-3-24-13-a.html#post5806236
If anyone has a copy of the .layout_version file, please post it or a link to it so I can check it out.
Thks
becomingx said:
Wow! This is very messy, no wonder you're trying to automate the cleanup.
Love the analogy here :laugh:
Looks to be `/data/.layout_version' [ref].
If we're going full steam ahead with the requirement of Busybox being installed, then this is pretty easy:
Code:
VAR="$(find */0 | sed -e 's!/0!!g' | sort -u)"
Sed is our Stream EDitor. The option `-e' tells sed that what follows is a sed script.
The script itself (s!/0!!g) breaks down like this:
* `s' means we are doing a substitution.
* The exclamation point (!) is just a delimeter; any character can be used. Typically a forward slash (/) is used, but since directory paths have forward slashes in them, we go with something else to make it easier to read. Otherwise we'd have to escape the forward slashes and it would look like this: s/\/0//g
* `/0' is what we are looking to match.
* Another exclamation point (!) indicates we are done with our "search" pattern and next begins our "replace" pattern.
* Since we are just looking to just get rid of all occurences of `/0', our replacement pattern is literally nothing.
* Another exclamation point (!) to indicate the end of the replacement pattern.
* Finally, `g' (for global) tells sed to keep applying the substitution until it no longer matches. This means a path like `/dir/0/0/0' gets each `/0' stripped away one at a time until there are none left.
The last command sorts our listing so that the `-u' (for unique) option can be applied to it, which removes duplicate lines.
The easiest way to do this is just check for the presence of a `0' dir at each level. If one exists then there's another level of nesting below us and we ignore the current level:
Code:
VAR1=""
for i in $(find */0); do
if [ ! -d $i/0 ]; then
VAR1="$VAR1 $i"
fi
done
Click to expand...
Click to collapse
Thanks again for your help and explanations. I find it amazing what a few (not so) simple lines of code can do with a complex task. Learning regex will take me until next year so for now I will go blindly forward with my plan. The next step would be to use BOTH of these variables in some call statement like [ ] to automate the process of backing up and moving data around. And combining it all into a script here it is below. Is the syntax in the main block I added correct?
Code:
#!/system/bin/sh
VAR="$(find */0 | sed -e 's!/0!!g' | sort -u)"
VAR1=""
for i in $(find */0); do
if [ ! -d $i/0 ]; then
VAR1="$VAR1 $i"
fi
done
for OBJ in $VAR and OBJ1 in $VAR1; do
#Backup $OBJ to SD
mkdir /Removable/MicroSD/$OBJ_BAK
cp -r /$OBJ /Removable/MicroSD/$OBJ_BAK
#Over write $OBJ with $OBJ1
cp -rf /$OBJ1 /$OBJ
#Merge backup and $OBJ
cp /Removable/MicroSD/$OBJ_BAK/*.* /$OBJ
#Remove $OBJ backup
rm -rf /Removable/MicroSD/$OBJ_BAK
done
elfaure said:
Wow! Thanks!! But now I'm suffering from cognitive dissonance (dis-sed-ance). I knew it would require either find, grep, awk, or sed but using those commands without having any grip on regex is nearly impossible. Can you point me to a good resource for learning Android regex or is any GNU/Linux resource valid? Is this a good one?
http://www.linux.org/article/view/introduction-to-regular-expressions-within-a-shell
Click to expand...
Click to collapse
lolz. Any resource on regexes should be applicable, just know that there are some small differences between the three main dialects (basic, extended, and perl). If you've got money to spend, Mastering Regular Expressions is considered *the* book to own on the subject. Also good is this pocket reference for sed and awk, which is a bit cheaper. I've used my copy so much the spine has completely gone out and it's just a collection of loose pages inside the cover now!
---------- Post added at 11:29 AM ---------- Previous post was at 11:00 AM ----------
_that said:
So does that mean you don't have any "0" directories in your /sdcard right now, and you are doing all this just for fun?
At least, after reading your additional material, I am now beginning to understand the problem.
However I don't understand how you get any useful output from "find */0" - I get either "No such file or directory", or a listing of thousands of files, depending on where I start this.
Click to expand...
Click to collapse
This is what I was hinting at back in post #3.
Using `find' like this, the `*/0' will get expanded by the shell if there happens to exist at least one `SOMEDIR/0', which will result in `find' printing out every file/directory/etc. that exists under any and every `SOMEDIR/0'. Otherwise, if none exist, it will get passed to the `find' command as literally `*/0', which is where the "No such file or directory" comes from.
The proper way to do this is using a `find' option like `-name', to limit scope of what gets returned. Again, though, one still has to be careful because we could still get search matches that we don't want. Something like this might be a good starting point:
Code:
find -name '0' -type d
@elfaure: Try populating your self-created `0' directories with some files and directories, the output of `find */0' should become a bit surprising.
But I don't know if you even need such powerful tools... Correct me if I am wrong, but your script should convert this structure:
/data/media/0
/data/media/0/stuff1
/data/media/0/0
/data/media/0/0/stuff2
/data/media/0/0/0
/data/media/0/0/0/stuff3
to
/data/media/0
/data/media/0/stuff1
/data/media/0/stuff2
/data/media/0/stuff3
?
So you'd basically need to run this inside /data/media/0:
1. If there is no subdirectory named "0", there is nothing to do, and we can exit
2. If there is a subdirectory named "0", move everything from inside one level out
2a. how do you want to handle conflicts, if the file or directory you want to move outside already exists in the parent directory?
3. go to step 1
Click to expand...
Click to collapse
Seems simple enough, but need a little time to think about it...
---------- Post added at 11:31 AM ---------- Previous post was at 11:29 AM ----------
elfaure said:
If anyone has a copy of the .layout_version file, please post it or a link to it so I can check it out.
Thks
Click to expand...
Click to collapse
Mine just contains the number "2", nothing else, not even a newline.
---------- Post added at 11:43 AM ---------- Previous post was at 11:31 AM ----------
elfaure said:
Thanks again for your help and explanations. I find it amazing what a few (not so) simple lines of code can do with a complex task. Learning regex will take me until next year so for now I will go blindly forward with my plan. The next step would be to use BOTH of these variables in some call statement like [ ] to automate the process of backing up and moving data around. And combining it all into a script here it is below. Is the syntax in the main block I added correct?
Code:
#!/system/bin/sh
VAR="$(find */0 | sed -e 's!/0!!g' | sort -u)"
VAR1=""
for i in $(find */0); do
if [ ! -d $i/0 ]; then
VAR1="$VAR1 $i"
fi
done
for OBJ in $VAR and OBJ1 in $VAR1; do
#Backup $OBJ to SD
mkdir /Removable/MicroSD/$OBJ_BAK
cp -r /$OBJ /Removable/MicroSD/$OBJ_BAK
#Over write $OBJ with $OBJ1
cp -rf /$OBJ1 /$OBJ
#Merge backup and $OBJ
cp /Removable/MicroSD/$OBJ_BAK/*.* /$OBJ
#Remove $OBJ backup
rm -rf /Removable/MicroSD/$OBJ_BAK
done
Click to expand...
Click to collapse
Negatory, good buddy. Everthing after the `in' keyword till the semicolon gets treated as part of the (one and only) list. This would be a good place for two parallel arrays, but I suspect that might be overkill. Need a little time to think...
P.S. Never go blindly forward when working as user `root'. Much too dangerous.
Hello all, im working on a app to transfer pictures from a camera to the phone and i would like some assistance. i got the camera connected after putting together some code reading around about it. but i can't or don't know how to transfer the files over. any assistance would be great. thanks!
inspector94 said:
Hello all, im working on a app to transfer pictures from a camera to the phone and i would like some assistance. i got the camera connected after putting together some code reading around about it. but i can't or don't know how to transfer the files over. any assistance would be great. thanks!
Click to expand...
Click to collapse
Use adb.exe
1) Set ADB to &PATH
2) Make a loop running "adb devices"
3) Parse out the devices when one is detected
4) Run adb -s "DEVICE-SERIAL" push /sdcard/ or pull /sdcard/ to send files across
Here is some Regex for java for parsing out adb devices if you have more than one device connected and want to add concurrency -
Code:
private static final String mRegex = "^([a-zA-Z0-9\\-]+)(\\s+)(device)";
Hope this helps,
Ryan
P.S - Some devices like the Kindles have non-standard file systems, you may want to make some logic around this if you plan to make it large scale; for example user input to the destination path using tree/ls mappings.
SkysRad said:
Use adb.exe
1) Set ADB to &PATH
2) Make a loop running "adb devices"
3) Parse out the devices when one is detected
4) Run adb -s "DEVICE-SERIAL" push /sdcard/ or pull /sdcard/ to send files across
Here is some Regex for java for parsing out adb devices if you have more than one device connected and want to add concurrency -
Code:
private static final String mRegex = "^([a-zA-Z0-9\\-]+)(\\s+)(device)";
Hope this helps,
Ryan
P.S - Some devices like the Kindles have non-standard file systems, you may want to make some logic around this if you plan to make it large scale; for example user input to the destination path using tree/ls mappings.
Click to expand...
Click to collapse
Thanks for the reply, i'll see if i can give this a try. Is there a way to do it with the mtp API? similar to how the android camera gallery works. cause else i have to redo the code. Thanks
Hello XDA comunity!
I want to share you a very useful tool that I have created, its name is: "kielyd".
What's keilyd?
Keilyd is a BASH daemon that constantly dumps the system's logs, warnings and other useful info in a easy-to-share "ZIP" with the main purpose to create a very informative and useful bug report.
keilyd is expansible and configurable. It's very easy to add new features because BASH interacts directly with the Linux/BSD commands in the Android System.
How does it works?
Keilyd is launched at device boot, with the "init.d support" and as root (see requeriments below). After launched, keilyd relaunches itself in a "daemon mode", that runs in the background, then, it dumps several information in the /sdcard and compress it in a single zip, so, the user can send that zip (in fact, are 2 zip's or more if you use the "snapshot" feature) to the developer of their ROM, to a forum or send it to a 3rd party developer and report a bug with all the system logs included.
Keilyd stores 2 zips: the first one is the current log (the "very last file") and the second one is the last log before keilyd were launched again.
If you won't enable de daemon, you can create a single zip when you use the "snapshot" feature, if you take two or more snapshots in the same minute, the ZIP will be the same, otherwise, one script per different minute will be created. The Snapshot feature can create as much zips as needed.
What info is collected?
That's a good question!
The info that is collected is (this applies ONLY for a vanilla copy of keilyd):
Kernel log (aka "dmesg") and other kernel info
Logcat (aka "logcat")
List of process running and memory statics
A copy of all tweaks in init.d (including keilyd itself)
A copy of build.prop and sysctl.conf
A copy of /proc/config.gz (if exists)
Some cpuinfo
Mounted filesystems and usage
How does this "daemon" can help me?
Developer:
Have you had the need to get some system log from a user?
Probably the answer is "yes", but, the user may have no idea how to do that, so, you have to give them a "step-by-step" guide of "how to get a system log in the terminal".
This simplifies the process, so, the user can report a bug with logs with no terminal (if the daemon is enabled) or with a single command that looks like
Code:
keilyd --snap
User:
Well, the complement of the developer... You can share "golden" information just by sending 2 zips files to the dev. Minimal Android knowledge is the only prerequisite if the daemon is working as spected.
This daemon is not a demon, but it can be very bad if not implemented as spected. Please, ensure that ALL THE COMMANDS ARE KNOWN and that THERE IS NO PORPRIETARY interference.
Why BASH?
A good idea is to create a daemon in C/C++ and share it as a "binary", BUT, as we are recolecting information from the device, the user must have to be concient of what are the script recollecting, if the user can't read bash, it can ask to a 3rd developer. You can feel safer because it's under the GPLv3 licence!
In addition, BASH run in (almost) any UNIX based OS, including but not limited to Apple OS X, GNU/Linux, Android, iPhone OS, etc...
This prevent a "blind trust" from the user, and give them facility of enable-disable the script, add features and so on.
Requeriments
- A rooted device
- Busybox (or run-parts)
- init.d support
- BASH 4.3.x ONLY SUPPORTED by the moment.
- Terminal emulator/HIDE]
How can I include it in my ROM?
Take a look of my repo in GitHub, it contains the last version of keilyd and the implementation during compilation in the Makefile.
If you are a ROM modder, download the latest version from above, include init.d and bash support in you ROM.
Or you can create a CWM/TWRP/Philz... flashable zip (it would be great!)
Users: how to use
When something goes weird, open the terminal emulator and write:
Code:
keilyd --snap
and send to the developer all zips in
Code:
/sdcard/keilyd/out
type
Code:
keilyd --help
and a helpful message will be printed.
If something goes wrong and you need to reboot the device (or the device reboots itself), just send both zips to the developer or the forum where you are asking help (This feature only works if the daemon is enabled).
Developers: how to use
You can add/remove things in order to fit your needs, the only thing that we ask you is that you write the whole script in BASH, without "blobs" that may private the user from know what is exacly logged and stored.
If some user sends you the zip, you will notice that the files are plain-text, and you can view it in any text editor.
TODO:
List all apps installed (system and user)
Secure "user-modificable" settings (Allmost DONE)
A graphical interface (an app) to manage the daemon
A flashable ZIP
Aports and sugestions are WELLCOME, please, coment in the box below
Install redistributable package:
If the avobe requeriments are meet...
1. Open this GItHub URL
Click in the "Download ZIP" icon in the right.
2. Put the zip in your phone
3. Unzip it.
4. If you want to use the script... A; otherwise, B.
A.
a. Open the terminal emulator.
b. cd to the folder that contains
Code:
install.sh
c. As root, type "
Code:
sh install.sh
" in the terminal.
B.
a. Open ES File Explorer
b. Enable root explorer and mount /system as rw
c. Copy
Code:
daemon/18keily
to
Code:
/etc/init.d
d. Tap and hold the
Code:
18keily
file and open the poperties
e. Change te permisions to
Code:
rwx r-x r-x
f. Copy the
Code:
redistrib/bash
file to
Code:
/system/xbin/bash
g. Change the permisions as above.
h. Symlink /etc/init.d/18keily to /system/xbin/keilyd
i. Reboot
NOTE:
Keilyd can be used in 2 ways:
1. As a daemon that does it job without human interaction, that constatantely and automaticaly creates 2 zips in the output directory.
2. As a Script that only needs to be called with the '--snap' parameter, but it requieres human interaction and does not constantely repeat the process.
Use the first one if you are a beta tester, or if you know that you ROM can hang sudently. Recomended for Unstable ROM'S.
Use the Last One if your ROM is stable or official, but you want to do a bug report.
You can allways use the second option if you use the first one.
By default, the daemon is disabled. You can enable if you change the "ENABLED" variable to "Y". I personally recommend values for "MINS" between 5 and 15.
BUGS REPORTS in the GITHUB REPOSITORY
Reserved
Hi there,
This looks like what I need, did you close the repo? Can you share a link to the script? I have a few devices, and I want to have some forensic information for troubleshooting (I.E. log all the warnings, errors and critical logs, + a snapshot every n minutes about the system usage). Right now I have to go where the device is, connect it with ADB and query the logs (most of the time, the logs were overwrited)... Thanks in advance.