File I/O: how can I disable auto-convert of the 2-byte sequence 0x0D, 0x0A into a single byte '\n'?

ruby-talk
1

Hi,
does somebody know, how to disable the auto-conversion of a 2-byte
sequence of 0x0D 0x0A into a single byte '\n'?
I want to process binary data, which can contain such byte sequences but
not idicating end-of-line as the file has a fix record length.
Please take the example below:

# input file "test.dat" contains 2 bytes: 0x0D and 0x0A
# the result file should contain 0x0A 0x0D in this example

dest = File.new("test.dat", "r")
a = dest.sysread(1) # should read 1 byte (0x0D), but does read both in
reality as '\n'
b = dest.sysread(1) # should read 1 byte (0x0A), but gets EOF/nil in
consequence of above
dest.close

dest = File.new("result.dat", "w")
dest.syswrite(b)
dest.syswrite(a)
dest.close

Best regards,
Frank

2

Have you tried switching to binary mode ?

dest.binmode

Cheers !

···

On Tue, Jun 03, 2003, Polscheit, Frank wrote:

input file “test.dat” contains 2 bytes: 0x0D and 0x0A

the result file should contain 0x0A 0x0D in this example

dest = File.new(“test.dat”, “r”)

–
Pierre Baillet
Avec un escalier prévu pour la montée, on reussi souvent à monter
plus bas qu’on ne serait descendu avec un escalier prévu pour la descente.
Devise Shadok

3

dest = File.new(“test.dat”, “rb”)
^

dest = File.new(“result.dat”, “wb”)
^

···

On Tue, 2003-06-03 at 05:55, Polscheit, Frank wrote:

–
Tom Felker

38% of statistics are made up on the spot.

4

file I/O: how can I disable auto-convert of the 2-byte sequence 0x0D, 0x0A into a single byte ‘\n’?

···

“Polscheit, Frank” frank.polscheit@gzs.de wrote:

Hi,

I want to process binary data, which can contain such byte sequences
but not indicating end-of-line as the file has a fix record length.

Best regards,
Frank

And here’s an example.

#---------
FIXREC_LEN = 10
FIXPAD_CHR = ‘!’

File.open(‘result.dat’, ‘wb’) do |fo|
File.open(‘test.dat’, ‘rb’) do |fi|
until fi.eof
dat = fi.read(FIXREC_LEN)
p [dat, dat[4], dat[5]]
dat_new = dat.gsub(/\r\n/, “\n\r”)
fo.write(dat_new << FIXPAD_CHR * (FIXREC_LEN-dat.size))
end
end
end # files close automatically for File.open {block}

puts ‘-’ * 30

File.open(‘result.dat’, ‘rb’) do |fi|
until fi.eof
dat = fi.read(FIXREC_LEN)
p [dat, dat[4], dat[5]]
end
end
#---------

Output:

[“ERST\r\nZWEI”, 13, 10]
[“T\r\nDRITT\r\n”, 82, 73]
[“VIERT\r\n”, 84, 13]

[“ERST\n\rZWEI”, 10, 13]
[“T\n\rDRITT\n\r”, 82, 73]
[“VIERT\n\r!!!”, 84, 10]

#File dumps:

test.dat

000: 45 52 53 54 0d 0a 5a 57 45 49 ERST…ZWEI

010: 54 0d 0a 44 52 49 54 54 0d 0a T…DRITT…

020: 56 49 45 52 54 0d 0a VIERT…

result.dat

000: 45 52 53 54 0a 0d 5a 57 45 49 ERST…ZWEI

010: 54 0a 0d 44 52 49 54 54 0a 0d T…DRITT…

020: 56 49 45 52 54 0a 0d 21 21 21 VIERT…!!!

daz

5

Thank you for helping.

For all, interesting in this question: alternativ to the method
“binmode”,
it works with File.new(“test.dat”, “rb”) as well, opening the binary-IO.

Cheers, Frank

···

-----Ursprüngliche Nachricht-----
Von: Pierre Baillet [mailto:oct@zoy.org]
Gesendet: Dienstag, 3. Juni 2003 13:38
An: ruby-talk@ruby-lang.org
Cc: rheinmainruby@eylerfamily.org
Betreff: Re: file I/O: how can I disable auto-convert of the 2-byte
sequence 0x0D, 0x0A into a single byte ‘\n’?

On Tue, Jun 03, 2003, Polscheit, Frank wrote:

input file “test.dat” contains 2 bytes: 0x0D and 0x0A

the result file should contain 0x0A 0x0D in this example

dest = File.new(“test.dat”, “r”)

Have you tried switching to binary mode ?

dest.binmode

Cheers !

Pierre Baillet
Avec un escalier prévu pour la montée, on reussi souvent à monter
plus bas qu’on ne serait descendu avec un escalier prévu pour la
descente.
Devise Shadok