Buffer to string

[Jari Williamsson]

I have a buffer (stored in a String class) with a NULL-terminated
C-style string inside, what's the most efficient approach to get that
text string up to the NULL?


Best regards,

Jari Williamsson

 

[elof]

Hi

One approach could be to use split and a regexp:

s = "hi"
s << 0.chr
s << "there"
p s
p s.split(/\000/).first

Kristian

 

[Paul Irofti]

I have a buffer (stored in a String class) with a NULL-terminated
C-style string inside, what's the most efficient approach to get that
text string up to the NULL?

$ irb
irb(main):001:0> s = "hello\0"
=> "hello\000"
irb(main):002:0> s.chomp("\0")
=> "hello"
irb(main):003:0>

See String#chomp(!) for more details.

 

[elof]

If you really only ever care about the first string then

s.match(/\000/).pre_match

is an alternative which looks like it could be less work for the machine
since it can stop after the first \000 rather than have to split the entire
string.

Are you looking at strings huge enough that efficiency at this level
matters?

Kristian

 

[Paul Irofti]

If you really only ever care about the first string then

s.match(/\000/).pre_match

is an alternative which looks like it could be less work for the machine
since it can stop after the first \000 rather than have to split the entire
string.

Are you looking at strings huge enough that efficiency at this level
matters?

I think the OP made it pretty clear: there is *one* C-Style string
inside a String Class buffer and wants to retrive it w/o the zero
termination.

 

[Tom M]

I have a buffer (stored in a String class) with a NULL-terminated
C-style string inside, what's the most efficient approach to get that
text string up to the NULL?

Best regards,

Jari Williamsson

Would buf.unpack("a*") work?

 

[Jari Williamsson]

$ irb
irb(main):001:0> s = "hello\0"
=> "hello\000"
irb(main):002:0> s.chomp("\0")
=> "hello"

This particular approach wouldn't work, since it can't be guaranteed
that every byte remaining in the buffer (after the termination) is NULL:
irb(main):001:0> s = "Hello\0\1\0"
=> "Hello\000\001\000"
irb(main):002:0> s.chomp("\0")
=> "Hello\000\001"

But the other suggestions work. Thanks!


Best regards,

Jari Williamsson

 

[Paul Irofti]

Would buf.unpack("a*") work?

Depends on the raw data contained within the C zero terminated string.
If its plain text "a*" wouldn't work, but "Z*" might.

For unpack I'd suggest seeing String#unpack for other formating and the
way to combine the different format markers in order to retrieve the
wanted data.

$ irb
irb(main):001:0> s = "Hello\0"
=> "Hello\000"
irb(main):002:0> s.unpack("a*")
=> ["Hello\000"]
irb(main):003:0> s.unpack("Z*")
=> ["Hello"]

 

[Paul Irofti]

This particular approach wouldn't work, since it can't be guaranteed
that every byte remaining in the buffer (after the termination) is NULL:
irb(main):001:0> s = "Hello\0\1\0"
=> "Hello\000\001\000"
irb(main):002:0> s.chomp("\0")
=> "Hello\000\001"

But the other suggestions work. Thanks!

I did not consider trailing garbage, sorry about that. I thought, from
your original question, that you only have a NULL terminated string and
not anything else.

 

[Paolo Bonzini]

Would buf.unpack("a*") work?

No, that's buf.unpack("Z*") that you want

irb(main):001:0> s = "hi"
=> "hi"
irb(main):002:0> s << 0.chr
=> "hi\000"
irb(main):003:0> s << "there"
=> "hi\000there"
irb(main):007:0> s.unpack("a*")
=> ["hi\000there"]
irb(main):014:0> s.unpack("Z*")
=> ["hi"]