struct.(un)pack and ASCIIZ strrings

[Sergey Dorofeev]

I can use string.unpack if string in struct uses fixed amount of bytes.
But is there some extension to struct modue, which allows to unpack
zero-terminated string, size of which is unknown?
E.g. such struct: long, long, some bytes (string), zero, short, short,
short.

 

[Terry Reedy]

I can use string.unpack if string in struct uses fixed amount of bytes.

I presume you mean struct.unpack(format, string). The string len must be
known when you call, but need not be fixed across multiple calls with
different strings.

But is there some extension to struct modue, which allows to unpack
zero-terminated string, size of which is unknown?
E.g. such struct: long, long, some bytes (string), zero, short,
short,short.

Size is easy to determine. Given the above and string s (untested code):
prelen = struct.calcsize('2l')
strlen = s.find('\0', prelen) - prelen
format = '2l %ds h c 3h' % strlen # c swallows null byte

Note that C structs can have only one variable-sized field and only at the
end. With that restriction, one could slice and unpack the fixed stuff and
then directly slice out the end string. (Again, untested)

format = 2l 3h' # for instance
prelen = struct.calcsize(format)
tup = struct.unpack(format, s[relen])
varstr = s[prelen, -1] # -1 chops off null byte

Terry J. Reedy

 

[John Machin]

I presume you mean struct.unpack(format, string). The string len must be
known when you call, but need not be fixed across multiple calls with
different strings.



Size is easy to determine. Given the above and string s (untested code):
prelen = struct.calcsize('2l')
strlen = s.find('\0', prelen) - prelen
format = '2l %ds h c 3h' % strlen # c swallows null byte

shouldn't this be '2l %ds c 3h'??

Note that C structs can have only one variable-sized field and only at the
end. With that restriction, one could slice and unpack the fixed stuff and
then directly slice out the end string. (Again, untested)

format = 2l 3h' # for instance
prelen = struct.calcsize(format)
tup = struct.unpack(format, s[relen])
varstr = s[prelen, -1] # -1 chops off null byte

Perhaps you meant varstr = s[prelen:-1]