bug#10030: grep: strange behavior with '-' in character class

[Erik Auerswald]

Hi,

On 11/12/2011 08:05 PM, Thomas Dignan wrote:

echo /lib64/ld-linux-x86-64.so.2 | grep -o '[a-zA-Z\/0-9\-\.]*'


You probably want to use something like

 echo /lib64/ld-linux-x86-64.so.2 | grep -o '[-a-zA-Z/0-9.]*'

Note: The '-' should be the first character inside a character class if 
it is not used to describe a range.


Erik

bug#10030: grep: strange behavior with '-' in character class

[Thomas Dignan]

Hello Coreutils Team,

When I use the command:

echo /lib64/ld-linux-x86-64.so.2 | grep -o '[a-zA-Z\/0-9\-\.]*'

The result is that I get a tokenized string where - has been replaced 
with \n


/lib64/ld
linux
x86
64.so.2

The expected result is:

/lib64/ld-linux-x86-64.so.2

When the character class is modified so that the - is not next to the 
number 9:


echo /lib64/ld-linux-x86-64.so.2 | grep -o '[a-zA-Z\/0-9\.\-]*'

/lib64/ld-linux-x86-64.so.2

The expected result is obtained.

I am aware - is used to specify a range, but it is escaped, should this 
be the proper behaviour?


Thanks,
Tom

bug#10030: grep: strange behavior with '-' in character class

[Thomas Dignan]

I have since learned that you are not supposed to escape things within 
brackets at all and simply position them so they have relevant meaning, 
and that dot does not need to be escaped inside brackets.


Not sure if my mail is still relevant.

On 11/12/2011 02:05 PM, Thomas Dignan wrote:

Hello Coreutils Team,

When I use the command:

echo /lib64/ld-linux-x86-64.so.2 | grep -o '[a-zA-Z\/0-9\-\.]*'

The result is that I get a tokenized string where - has been replaced 
with \n


/lib64/ld
linux
x86
64.so.2

The expected result is:

/lib64/ld-linux-x86-64.so.2

When the character class is modified so that the - is not next to the 
number 9:


echo /lib64/ld-linux-x86-64.so.2 | grep -o '[a-zA-Z\/0-9\.\-]*'

/lib64/ld-linux-x86-64.so.2

The expected result is obtained.

I am aware - is used to specify a range, but it is escaped, should 
this be the proper behaviour?


Thanks,
Tom

bug#10030: grep: strange behavior with '-' in character class

[Bob Proulx]

tags 10030 + notabug
close 10030
thanks

Thomas Dignan wrote:
 I have since learned that you are not supposed to escape things
 within brackets at all and simply position them so they have
 relevant meaning, and that dot does not need to be escaped inside
 brackets.

That's right.  You have it.

 Not sure if my mail is still relevant.

Well...  Not really.  But mostly because you sent your bug report to
bug-coreutils instead of bug-grep.  Wrong bug list!

Since you have it figured out and it is the wrong list I am going to
go ahead and mark the bug as closed.  You can follow up with bug-grep
if you have another report.

Bob