h3	#target .P, .81 and .P81
	p, 81, p81
	
p.magenta <i>this format is not yet well tested. If there are problems please send an email to <a href="mailto:kio@little-bat.de">Kio</a>.</i>

p	This creates a tape file for use with a ZX81 emulator. The tape file will be loaded to the ram start + 9 at address $4009. 

p	"p" can be used instead of "81". This will create a target file with extension "p" or "81" respectively, which are fully identical. ".p81" files also include the program name which must be written using the ZX81 character set.

p	Code may be stored in one or more #code segments. However, the first segment must be at least 0x403B - 0x4009 bytes long and contain the system variables. (except for the first 9 bytes which are not saved.)



pre	#target	81
	#charset ZX81
	#code <name>,<start>,<len>
		...

p	or:

pre	#target	p81
	#charset ZX81
	#code	_PROGNAME
		dm	"progname" | $80		; zasm converts the characters to charset ZX81!
	#code 	_RAM, 0x4009, _ram_end - 0x4009
		...

p	See the ZX81 .p template file: <a target="image" href="../Examples/template_p.asm">template_p.asm</a> 
	
p	The tape data is always loaded to $4009, so this is the only choice for the <start> address in the #code directive. The <size> may be any value up to a maximum of $C000 minus $09, but most ZX81 had at most 16 kB ram if extended with an external memory expansion, only 1 kB if none. (The TS1000 had whoopy 2k!)

p	The ram starts with system variables from $4009 to $403C, which must be set to proper values. The variable E_LINE at the address $4014, which contains the end of ram address, is checked by the assembler. 

P	This file also contains the screen file, which basically means, that the video memory must contain 25 HALT opcodes if the file was saved in SLOW mode.

h6	Here are few important variables:

pre	$4014	defines the end address (used to calculate the file length)
	$4029	points to the next executed (autostarted) BASIC line
	$403B	bit 6 indicates if program runs in SLOW or FAST mode 
	$403C++	may be misused for whatever purpose

p	zasm has a convenient assembler directive to translate from the ascii (or utf-8) characters in your source file to the non-ascii character set of the ZX81:

pre	#charset ZX81

p	You can include c sources in your assembler file, but there are two pitfalls:

p	You can use the character set translation of zasm to translate strings in your c source as well. But unluckily character literals are not exported as character literals by the c compiler sdcc but as their ascii code and zasm has no chance to detect this. So character literals in your c source are not translated.

p	Second, character 0x00 is used in c sources as a string end indicator, but it is also a valid character for the ZX81: the space. So you'll have to think of a way to work around this problem.