World: r4wp

Sorry, messed that up

Fixed it

Now I get this beautiful output:

¢÷¶¢÷¶ÿÿÿÿð£÷¶ð£÷¶ 20
¢÷¶¢÷¶ÿÿÿÿð£÷¶ð£÷¶/img00.bin 30

¢÷¶¢÷¶/img00out.bin

*** Runtime Error 1: access violation
*** at: B6EC5C0Ch

I must be doing something wrong...

At least the lengths are showing up in the right place...

It seems that the 36 is one too high. Does it also print that on 
Windows?

Here's Windows:

¢÷¶¢÷¶ÿÿÿÿð£÷¶ð£÷¶ 20
¢÷¶¢÷¶ÿÿÿÿð£÷¶ð£÷¶/img00.bin 30

¢÷¶¢÷¶/img00out.bin

*** Runtime Error 1: access violation
*** at: B6EC5C0Ch

Sorry, here's Windows:

+?3+?3 6
+?3+?3/img00.bin 16
x?3
8?3/img00out.bin

*** Runtime Error 1: access violation
*** at: 77C47AB4h

Did you reverse copy-string file3 file1 ?

No.  That would probably be important.

But with it reversed, it looks strange.  It's the opposite of the 
order of 'append-string, 'find-string and all those other operations 
now.

Do I also reverse 'copy-string-part?

Yes, I think it's a more natural order. You copy or move from to

Here on Intel Linux I get:

to-process/2013-06-29-20-18-55.h264 35
to-process/2013-06-29-20-18-55.h264/img00.bin 45

Without the reversal, you were overwriting file1 with random memory 
from file3

Yes, I can see that now.

Works fine on Windows again, but on Linux-ARM I'm still getting messed 
up:

	 36process/2013-06-29-20-19-37.h264
	/img00.bin 4613-06-29-20-19-37.h264

Everything points to null endings lacking

Is that on Arch Linux?

No, I'm using Raspbian Wheezy right now.

Odd, that's probably just the GNU C library

Based on Debian, so probably.

Try adding an explicit null marker after each append

OK.  Currently 42C (108F) outside my house.  My small window air 
conditioner can't keep up.  My fingers are slipping on the keys.

Wow, my hard disk runs that hot

It was 44C today but it is starting to cool off.

My official C book doesn't mention null ending for strcat, but it 
does for all the other string functions, so it would be a very malevolent 
C library that doesn't implement it

Is this a good way to add that null marker?

	tmp: make-c-string 1
	append-string file1 first-line
	tmp: file1 + length? file1
	tmp/1: #"^(00)"

No, you have to ask for the length before the append, because LENGTH? 
depends on the marker

OK.

Here's my code now...still not working:

	tmp: make-c-string 1
	file1: make-c-string 128
	copy-string "to-process/" file1 ;reversed
	print-line [length? file1 " " length? first-line]
	file1len: (length? file1) + (length? first-line)
	append-string file1 first-line
	tmp: file1 + file1len
	tmp/1: #"^(00)"

	print-line [file1 " " length? file1]
	append-string file1 "/img00.bin"
	print-line [file1 " " length? file1]

Here's the output

	11 25
	 36process/2013-06-29-20-20-09.h264
	/img00.bin 4613-06-29-20-20-09.h264

length? is correct

Aha, that means there's an extra character at the end of your first-line

Yes.

Judging by the printing, that must be a CR

So you probably copied a Windows text file to Linux

So the problem is in the only piece of the code you didn't show :-)

I could have pasted the whole program in here, but I thought that 
might have been excessive! :-)

Sorry for the false alarm and all the trouble.

Is first-line a literal or does it come from a text file?

Here's the troubling bit of code:

	dirs: make-c-string 1024
	dirs: read-file "to-process/dirs.txt"

	eol: make-c-string 1024
	eol: find-char dirs #"^/" ;Finds the first end-of-line character

 line-len: as-integer eol - dirs + 1 ;Add one byte for the end-of-string 
 character
	first-line: make-c-string line-len
	copy-string-part dirs first-line as-integer eol - dirs ;reversed
	first-line/line-len: #"^(00)"

Shouldn't 'find-char dirs #"^/" return the first newline byte?

Yes, you're looking for the LF, so you're including the CR

On Windows, the C library that executes the read-file converts DOS 
newlines to Unix LF format. I've had trouble with that before in 
the binding

The C library on Unix just assumes it's a native file, so if there 
are extra CRs in there, it doesn't touch them

It's best to standardise on LF format, even on Windows, with a capable 
editor

The dirs.txt file is being written to the Pi using Rebol on Windows.

Yes, it uses local Windows convention

You must do the conversion somewhere in the chain

So, I'd have to form the file in Rebol, convert it to binary (in 
Rebol) and remove the CRs from the binary, then use write/binary 
to write it out.