Archive for the ‘howtos’ Category.

Migrating multi-project Subversion repositories to Git

So I finally gave in and submitted to the version control Gods, Git is the way to go.  Of course it would happen that Git repositories are typically a single-project-per-repo situation, and I have a couple globbed together Subversion repositories with many projects in them.  How do I separate out my histories from the individual projects, maintain tags and branches, and not lose that precious history that nobody really ever looks at anyway?

Well, luckily there is a nifty script called svn2git that gets us the bulk work of migrating history of a single Subversion repo to a new Git repo.  Coupled with a few admin tools that come with Subversion, we got it covered.

The current situation

I have a couple repositories, looking like the following:


The desired situation

I’d like them all separated, with all history!!!  Resulting in the following structure:


Step 1 – Dump your projects from the Subversion repository

Subversion ships with a tool, svnadmin, that will help dump a repository.  It must operate on the local filesystem of the Subversion repository, not a working copy.  As a result, it only dumps full repositories, hence the problem in general.

Be not afraid though!  Subversion also ships with svndumpfilter, the key to our salvation.

For my example above, to dump project1, project2, project3 and project4 into separated Subversion repositories, I would:

mkdir svn-convert && cd svn-convert
svnadmin create project1.svn
svnadmin create project2.svn
svnadmin create project3.svn
svnadmin create project4.svn
svnadmin dump /var/lib/svn/repo1 | svndumpfilter include project1 | svnadmin load project1.svn
svnadmin dump /var/lib/svn/repo1 | svndumpfilter include project2 | svnadmin load project2.svn
svnadmin dump /var/lib/svn/repo2 | svndumpfilter include project3 | svnadmin load project3.svn
svnadmin dump /var/lib/svn/repo2 | svndumpfilter include project4 | svnadmin load project4.svn

And viola, separated Subversion repositories.

Step 2 – Convert via svn2git

First, download and install svn2git, it’s a Ruby script, the instructions are good, it’ll work for you.

Next, create a text file, ‘authors.txt’ with contents like the following:

name1 = Full Name <>
name2 = Another Name <>

Substituting the names of users who have committed to the Subversion repositories of course. This will map the Subversion users to Git authors.

Next, we make a place for the new repository and run the conversion utility.  It will init the Git repository and create the necessary history.

mkdir project1.git && cd project1.git
svn2git file:///home/nick/svn-convert/project1.svn \
    --trunk project1/trunk --branches project1/branches \
    --tags project1/tags --authors ../authors.txt

Repeated for each repository with appropriate substitutions, and you’ve got yourself some Git repositories.

Quick verification should show you you’re in business:

$ git tag
$ git branch
* master

All that is left is to put them in a safe place and start cloning!

Restoring from a Samba-based Time Machine backup (kinda)

Panicking about not being about to find or mount your Samba share you’ve been blissfully backing up to over your network?  Trying to restore to a new hard drive using the Leopard boot DVD?

I recently had the pleasure of a hard disk crash on my MacBook, and only a month earlier had started backing up to a Samba-share via Time Machine.  I had the “how to restore” question nagging in the back of my head when I set it all up but I figured someone had it figured out, otherwise why would so many articles exist to show you how to set it up?

I’ve been trying to find out how people using the TMShowUnsupportedNetworkVolumes hack to use Time Machine on smbfs shares have been restoring and it would seem that the answer is “they aren’t”.  That, or they’re simply using it as an “oops” fixer, restoring a file here and there.

When booting from the Leopard DVD, firing up the terminal and attempting to mount the share, the following delightful message shows up:

mount_smbfs --> mount_smbfs: failed to load the smb library: Unknown error: 1102

Searching for that was even more disappointing.  Other people running into the issue, no solutions.  It seems smb support is just not available on the boot DVD.

I ran into a possible solution, copy the Time Machine sparse bundle onto a removable hard disk, and hook it up to the laptop. Unfortunately all my external storage is formatted ReisferFS or ext3, neither are supported filesystems, and I didn’t feel like changing one just to fix this.

So in comes the hack.  Luckily the Samba share is on an Ubuntu 9.04 (Jaunty) server, so adding support for something Apple does support on the boot DVD is tragically easy. This is a fairly specific solution, but variations on it will work for many different servers.

Enter AFP

Looking through the other available mount applications, we also have mount_afp available.  This mounts Apple Filing Protocol-based shares, and it works too, bonus!

So it boils down to enabling AFP on the server and sharing the same volume via AFP.  AFP on Linux (BSD, etc) is supplied by netatalk, and here’s a step-by-step of how I wrapped it all up.

On the server:

  • sudo aptitude install netatalk
  • Edit /etc/netatalk/AppleVolumes.default
  • Add entry for the volume, such as:
    • /mnt/time_machine "tmachine"
  • Save the file
  • sudo /etc/init.d/netatalk restart

On the Mac:

  • Boot from the Leopard install DVD
  • Enable Airport (if on WiFi), join your network
  • From the menu bar, select Utilities -> Terminal
  • Navigate to /Volumes
  • Create a new mount point for the Time Machine volume
    • mkdir /Volumes/tmachine
  • Mount the AFP share on the new point (details)
    • mount -t afp afp://username:password@server.hostname/tmachine /Volumes/tmachine
  • Quit Terminal
  • Back at the main menu bar, select Utilities -> Restore System From Backup…
  • You should see your Time Machine backup volume listed
  • Select it, and select the date from which you wish to restore
  • Wait a considerable amount of time for it to determine the space needed
  • Enjoy the hours and hours of restore time!


The basic AFP installation added to the server is likely pretty insecure, I purged it as soon as the restore completed.  Read this for a more formal treatment on setting up an AFP server on Linux.  It is likely that the real solution is to stop suggesting people use Samba as a file server for Time Machine backups, instead switching to AFP altogether.


Just recently came back to this for a Snow Leopard restore, the directions are unchanged.  The nice bonus is there have apparently been some optimizations during the restore, and an initial space calculation was practically instant.  Nice.

Conditional Jumping in PIC16 Assembly

PIC Microcontrollers have a funky way of handling conditionals. I’d like to present a set of macros I’ve made to make this easier to use, as well as explain the basics behind the technique in general.

Most MCU’s I’ve worked with before PICs had nice simple conditional statements… The mnemonic was usually to the effect of “branch if x to destination”. Not so on the PICs.

On a PIC, we have to do an operation between the data to test, then either skip the next instruction or not, based on the results of this test.

Take the instruction “BTFSS” meaning “Bit-test F, Skip if Set”. This instruction takes a register and a bit number, and will jump over the next instruction if that bit number in the register is set. For example:

We have 2 registers, REG1 and REG2. We want to know if the value in REG1 equals the value in REG2. A quick, simple way to do this is first to load one into W, then subtract the other from it, with the result going in to W as not to destroy one of the variables. This sets the “ZERO” bit in the STATUS register to either 1 if the operation resulted in a 0, or 0 otherwise. We know if the operation resulted in 0 the two values are equal, so we can code like so:

  movf REG1, w
  subwf REG2, w
  btfss STATUS, Z
    goto NOTEQUAL
  ; they were equal
  ; they weren't equal

So we can test for equality.

We can also macro-ize it for ease of use…

BEQ macro REG1, REG2, DEST	; branch if REG1 == REG2
  movf REG2, W			; W &lt;- REG2
  subwf REG1, W			; W &lt;- REG1 - REG2
  btfsc STATUS, Z		; if result was nonzero: skip out
    goto DEST			; otherwise jump

This gives us a macro we can call with 2 registers and a destination, and have it jump there if the condition ends up being true, and just pass on through if it’s not true. Much, much easier to use.

The other tests: inequality, less than, greater than, less than or equal to, greater than or equal to and so on follow a similar pattern. They are all covered in the “” file I use quite often. I have both register-register comparisons and register-literal comparisons in there. Feel free to grab a copy and use it in your next project.


PIC buttons (interrupt-based)



Previously, in PIC buttons (polling) we saw how to poll for the state of a line connect to a button, that is all fine and good but really that is not the best way to do them. The “real” way to interface with external components like that is through interrupts, a slick feature.

Interrupts provide you with lots of freedom in your code. They allow you to sit back, relax, and be told when an event occurs, and not be forced to sit and wait for it to happen.

For this program, the schematic and circuit are practically the same, the only thing that changed location is the button.

Instead of looping over and over again, we simply wait, using a goto $ we are essentially goto’ing the same address over and over, “goto here, goto here, goto here…” ad nauseum. A common technique is also to use a SLEEP command, which puts the PIC in a low power mode and halts the program counter. Same effect to the user though. You of course could do ‘real’ work too instead of just burn cycles.

Once the button is pressed and released, the PIC will generate an interrupt, forcing the program to goto memory location 4. This is labeled in the code as ISR (Interrupt Service Routine).

For this program we are using the RB0/INT Interrupt. This interrupt occurs when there is a low-high change in PORTB,0. It can also be configured for high-low as well.

To enable this, we set INTCON,INTE. This bit says we want to know if a change occurs. To enable interrupts in general, we must then set INTCON,GIE. This lets all enabled interrupts occur.

We then wait for the interrupt. Once it occurs, GIE is automatically cleared so we can’t have them inside each other, and blink the LED a couple times. We then clear the interrupt flag, saying we’ve handled the interrupt (INTCON,INTF). We then re-enable interrupts and return from the interrupt: retfie.

It should be noted that if we were using more than one interrupt type, we would have needed to check the flag bits to find out which one interrupted us. We then handle it, and clear its flag. The PIC is somewhat crippled in this manner. Any and all interrupts generated and thrown into address 4 and “we” have to figure out which one occured. Many higher end MCU’s will have a table of addresses to jump to for each particular interrupt type, we then code in each location the correct routine and the processor knows which to call based on what happens.


PIC buttons (polling)


Polling for button input, how useful! This is pretty brief and gives a good idea how to let buttons control your programs execution.

In this tutorial I’ve switched from using an oscillator to using a crystal. This changes the design a bit. Using the 2 OSC pins, OSC1/OSC2 (CLKOUT/CLKIN) they hook to the crystal in parallel. Then the two sides of the crystal are connected to ground via two capacitors, in this case 18pF. The speed and capacitance needed varies, and can be seen in most any PIC MCU datasheet as to how to lay it out. It will also be in my schematic.

Please note this is essentially NO DIFFERENT than the oscillator. It is simply a different means of providing a clock signal. You can swap in the oscillator to the normal CLKIN pin and it will work just fine.

For this program, we’ll start the processor, and wait for a button to be pressed. After it has been pressed we’ll essentially execute the LED blinker from before.

Polling is extremely simple. Polling is the act of checking something for a certain state to occur. In our case, we poll PORTB,4 waiting for it to go low. Once it does, we continue on and start blinking.

Circuit Image

Circuit Image

Another topic that needs to be covered is that of Pull-Up resistors. In order for a button to be able to change from 0 to 1 (Ground to 5V) we need a way to protect everything from a short circuit. We do this with a Pull-Up (or Pull-Down in some cases) resistor. For our case, we connect PORTB,4 to +5V via a 10k resistor. So when we fire up the program, the PIC sees PORTB,4 as HIGH. We then connect our button to PORTB,4 on one side, and Ground on the other. Now, when we press the button, PORTB,4 is connected to ground and is now LOW. A short circuit between +5 and Ground would occur if not for the resistor, which being 10k limits the current to a measly 500uA, nothing to worry about. Once the button is released, Ground is disconnected and the pin returns to a HIGH state.

Reading the code, you may also notice the _BANK macro has changed a bit. I’ve simply modified it to encompass all possible BANK configs instead of using 4 different macros for each. Reading through it can give you a little insight as to how the conditional assembler works. It’s a bit like C/C++ and a bit like BASIC. Quite nice and handy.

A final new bit is the _MCLRE_OFF in the __CONFIG line at the beginning. This frees us from having to pull _MCLR high to keep from a RESET condition. Just keeps our parts count down. :)

Now for the new section of code.

  btfsc PORTB,4
    goto $-1

That’s polling, yup, that’s it. All it is doing is a bit test on PORTB,4, waiting for it to become clear (Ground). The goto line is telling it to go 1 instruction back ($ means the address of the current instruction, $-1 means one before, the btfsc). Once that pin becomes 0 it skips the loop forcing it to check, and lets it continue on down to the blinker!


PIC LED blinker (busy-wait)

LED blinker #1 circuit schematic


Beyond all doubt, the #1 beginning program in microcontrollers is the LED blinker. It’s super simple, and teaches the concept of pin voltages and busy-waits. Here is a busy-wait LED blinker program, and a walkthrough building it in MPLab.

First, the delay. This is a busy-wait delay program, busy-wait means you just burn instruction cycles for the delay, keeping the MCU “busy”. There’s a tiny bit of math behind them. First, the clock speed is 20MHz, the instruction frequency is (clock/4) so our instructions are executing at 5MHz. This gives us a period of 200ns per cycle.

I created two delay functions, for versatility, and for LCD stuff which will come later but it’s handy here. One is a “DELAY_US” which will delay a specified amount of microseconds. This is done by wasting 5 cycles (5*200ns = 1us) a specified amount of times (less than or equal to 255us, since I only made an 8-bit “delay” variable). We can learn the cycle times from the data sheet, and make it work from there.

Next, I created a “DELAY_MS” which delays a specified amount of milliseconds. Same 8-bit limitation of max 255ms, but that’s enough to have fun with… It simply calls DELAY_US a few times with specified amounts of delay, adding up to 1000us (=1ms) and repeats as many times as we tell it to.

These delays are used to make the LED blinking visible, otherwise it would blink faster than we could see (if at all, it takes a cycle or two for the pin to change state so it might not even change if we just toggle it every other cycle).

Here’s a little more PIC architecture information. The PIC’s data registers are broken into “banks” (bank 0, 1, 2, 3). Meaning you cannot get at them all at the same time, although some are mapped to all banks so you CAN get at them, important ones. We usually hang out in bank 0… This usually isn’t a problem, just something you need to remember. The data sheet illustrates it pretty well. When it comes up I’ll clarify things about it.

Also, for the I/O pins. Since they’re bi-directional, you need to choose which direction to set them to. Input or output. This is done by setting the tri-state register for the given port, for example PORTA’s tris register is called TRISA (not tough!). You set the direction of a specific bit, by setting the bit of the TRIS register to either 0 or 1, 0 meaning OUTPUT and 1 meaning INPUT. Not tough to remember: 0 = Out and 1 = In.

Source Code

Ok, first you need to make sure you have MPLab from

Once you have MPLab, download the LED Blinker (busy-wait) source code.

MPLab is a nice IDE, you’ll need to create a “project” and then pick your chip, and add the asm file to it (called a “node”). All code is going to be indented 2 spaces, labels will not be indented at all, assembler directives are either 1 or 2 spaces in…

First few lines are kinda simple, the title directive just sets a title for your project…

 title  "LED Blinker Tutorial 1"

LIST R = DEC sets the default “radix” for the program, meaning the number base. So if I put 100 in a line somewhere, it means 100 DECIMAL. If I changed it to LIST R = HEX, then if I put 100 in somewhere, it means 0x100, TOTALLY different. I find DEC easier to work with, and you can still use 0x whatever and it means HEX so you get the best of both worlds.

Next we INCLUDE the “” which will give us nice little names for our registers so we don’t need to remember their addresses, how nice of Microchip.

Then the __CONFIG line, arguably the ugliest line of code while still being fairly simple. You can see the list of __CONFIG’s in the “” file, they’re simply setting the configuration of certain chip features, the ones I have listed do this: _CP_OFF copyprotect OFF, you can set the copyprotect bits to make your chip unreadable, this is only good for a production product, don’t use it. _WDT_OFF, watchdog timer off, watchdog timer is there for mission-critical applications, it’s constantly running and needs to be reset constantly, if it isn’t reset it will reset the chip (assuming your program has locked up), we don’t need it here. _HS_OSC specifies a high speed oscillator, it’s in the data sheet. _PWRTE_ON turns on the power up timer delay, it’ll make the MCU wait a bit before executing to make sure voltage is stable, oscillator is stable, etc… _LVP_OFF – low voltage programming, dun’need it… _MCLRE_ON makes us hold MCLR high, instead of letting the chip do it…

; Variable declarations
 CBLOCK 0x20
DELAY,DELAYTMP			; delay function variables...

Now to the variable declarations, the CBLOCK directive lets us just list out our variable names, and the assembler will assign addresses for us, this is handy. The 0x20 is the starting address of general-purpose registers in BANK 0. We list em out, then end it with ENDC.

; Macro declarations
BANK0 macro           		; Switch to BANK0
  bcf STATUS,RP1
  bcf STATUS,RP0
BANK1 macro           		; Switch to BANK1
  bcf STATUS,RP1
  bsf STATUS,RP0
  movlw TIME
  movwf DELAY
  call DELAY_MS
  movlw TIME
  movwf DELAY
  call DELAY_US

Next the Macro’s… Macro’s are one of the coolest features of MPLab, it’s kinda like an inline C function, and kinda like a #define. When called, the code is dumped into where it was called from, but you can use variables in it, and even arguments to customize it’s compiling, stuff I’ll show in later programs.

Anyways, you make one by saying: NAME macro in the first column, then code. End it w/a endm. The ones I have are fairly simple and re-usable. BANK0 sets the bank bits to get us into BANK 0, go figure. BANK1 sets them to get us into BANK1, crazy!

DELAY_MILLI takes the TIME argument and loads it into W, next it moves W to the register labeled DELAY. Then it calls our illustrious DELAY_MS function which will be explained in detail down below… DELAY_MICRO does the same damn thing with DELAY_US!

; Program code
 org 0
  goto MAIN
 org 4
  ; interrupt handler

PAGE is a pagebreak for printing, though it doesn’t work for me. :)

org 0 tells the assembler to start assembly at address 0, our first instruction is to jump (goto) the label MAIN.

org 4 starts us in the Interrupt address of PIC’s… This is kind of strange and will be explained later on, for now just accept it as fact…

And if you’re wondering if we lost space for 3 instructions between 0 and 4, you’re right. :)

; Subroutines
DELAY_US			; busy wait of DELAY us
				; 200ns instruction period assumed
  nop				; (1)
  nop				; (2)
  decfsz DELAY,f		; test DELAY count (3)
    goto DELAY_US		; loop if not done (4,5)
  return			; gtfo (4,5)
DELAY_MS			; busy wait of DELAY ms
				; dependant upon DELAY_US being accurate
  movf DELAY,w
  movwf DELAYTMP		; save DELAY time
DELAY_MS_LOOP			; inner loop
  movlw 245			; load 245 (1)
  movwf DELAY			; into DELAY (2)
  call DELAY_US			; wait 245us (3-249)
  movlw 245			; load 245 (250)
  movwf DELAY			; into DELAY (251)
  call DELAY_US			; wait 245us (252-498)
  movlw 245			; load 245 (499)
  movwf DELAY			; into DELAY (500)
  call DELAY_US			; wait 245us (501-747)
  movlw 246			; load 246 (748)
  movwf DELAY			; into DELAY (749)
  call DELAY_US			; wait 246us (750-997)
  decfsz DELAYTMP,f		; test DELAYTMP count (998)
    goto DELAY_MS_LOOP		; loop if not done (999,1000)
  return			; gtfo (999,1000)

Next we have our subroutines, the delays, I have these before the main lines of code just out of habit, it’s not required.

DELAY_US… Pretty simple really, we start out by burning 2 cycles, so we’ve waited 400ns so far, next we decrement our counter, and test if it’s zero, if it isn’t, we goto DELAY_US, looping again, if not, we return. The test itself takes one cycle (600ns so far) and either the goto or return take 2 cycles (1000ns = 1us) so we have our microsecond delay!

DELAY_MS works on the same principle, it’s accurate enough for this! :)

; Mainline of code
  bcf TRISA,2			; PORT A, bit 2 is our output pin.
  bsf PORTA,2  			; set her.
  DELAY_MILLI 250		; wait 1/4 sec
  bcf PORTA,2			; clear her!
  DELAY_MILLI 250		; wait 1/4 sec!!!!
  goto LOOP_BEGIN		; forever... :o

MAIN is our label for the beginning of the code, jumped to by the first line up @ org 0. BANK1 gets us into BANK 1 so we can set our bit direction, we clear bit 2 of TRISA making bit 2 of PORTA our output pin, then we hop back to BANK 0…

Then our introductory programming teachers worst nightmare, a purposely created infinite loop. We label the beginning, then set our pin high, shutting off the LED (as you’ll see in the wiring diagram). We wait 250ms via our handy delay function, then clear the bit, turning the LED on, we wait again and loop ad nauseum.

‘end’ tells the assembler to give up…


Alright, so we have our program, we run the assembler by clicking the weird funnel icon or by going to Project -> Build Node (or All). It’ll crunch and come up with no errors (of course).  Then just toss it into the programmer and feed the chip your tasty code.


The assembled circuit

The assembled circuit

Wire up the circuit as in the schematic at the top of the page. Hopefully it illustrates to you why the LED is on when the bit is off, and off when the bit is on… The LED is a typical ~2V yellow LED… Wired up it should look something like the image.


Hook up +5V and Gnd, and fire it up! If everything is set up correctly you’ll get a steady blinking LED!

Something to try

Connect a momentary switch to the _MCLR line, wired to Ground on the other side. Pushing the button will reset the chip, releasing it will start it over from the beginning of the program, of course it will do the same stuff, but this demonstrates how the reset buttons work. I also highly recommend changing the code to add a more interesting blink pattern, longer/shorter delays, and other stuff to get used to modifying code…


Basic PIC16 hardware setup

Your basic PIC16 microcontroller can’t hold down the fort by itself, it needs a little help from a few components.



I use a prototyping breadboard, the white kind you can just plug & unplug stuff to all day long. It makes life prototyping a whole lot easier.

Ok, now on to the PIC stuff. First and foremost, you need an oscillator, the oscillator is what keeps the PIC moving, it provides the clock signal for the chip. For our 16F628, I used a 20MHz oscillator in a “half-can” package. Many people like to use crystals and capacitors, etc, I use them on simpler designs, but they’re a bit harder to understand at first.

These oscillators are powered by +5V, and the output goes into the pin labeled OSC1/CLKIN on the 16F628. You can get these oscillators at Digi-Key (part#: CTX169-ND).

The other key part is the reset line. We need to hold this line HIGH (+5V) and drop to it 0V to reset the chip, which we won’t need for a while. To tie it high you use a “pull-up” resistor (10k is good for this) to the +5V line, connected to the _MCLR pin of the chip (reset pin). I like to use the outside rails for Ground, and inside rails for +5V as you’ll see in the pictures… This isn’t *required* as our chip is smart enough to pull up it’s own _MCLR pin, but for the first few projects we’ll explicitly do it so you get used to it. :)



Also connect Vss and Vdd to Gnd and +5 respectively to power the PIC.

After wiring these parts up, your breadboard should look something like this:

Basic PIC16 wired up

Basic PIC16 wired up

We also need a power supply. I use an old AT power supply, taking the red and black lines for +5 and Gnd respectively. You can use a 5V bench supply, batteries, wall-wart, anything 5V (but make sure it’s regulated).

Another smart part to use is a bypass capacitor, this lets the PIC draw power from the cap instead of the power supply during an increased power draw time, these are usually brief, but the power is needed quickly and a voltage drop would be bad, so you place a small, fast capacitor nearby the IC to provide the instant power and keep everyone happy. Typical values are a .1uF Tantalum capacitor.

Also handy are a pair of IC chip pullers, available in most little toolkits, they’ll help you avoid damaging the chip as you pull it out of the breadboard…