This is the mail archive of the cygwin mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [ANNOUNCEMENT] Updated [test]: grep-3.0-2

On 2017-02-15 21:25, Steven Penny wrote:
> On Tue, 14 Feb 2017 10:17:25, "Eric Blake (cygwin)" wrote:
>> This build modifies the behavior of grep to no longer force text
>> mode on binary-mounted file descriptors.
> Works, thanks:
>     $ printf 'hello world\r\n' | grep . | od -tcx1
>     0000000   h   e   l   l   o       w   o   r   l   d  \r  \n
>              68  65  6c  6c  6f  20  77  6f  72  6c  64  0d  0a
>> Since this includes pipelines by default, this means that if you
>> pipe text data through a pipeline (such as the output of a windows
>> program), you may need to insert a call to d2u to sanitize your
>> input before passing it to grep.
> This is certainly a good way to do it, but for more portable solution
> use tr:
>     $ printf 'hello world\r\n' | tr -d '\r' | od -tcx1
>     0000000   h   e   l   l   o       w   o   r   l   d  \n
>              68  65  6c  6c  6f  20  77  6f  72  6c  64  0a

Until you hit Windows text files using <line>\r<overprint>\r\n [grumble], 
when sed is better: ...| sed 's/\r$//' |... or to remove bold/highlight 
doubled line overprint only: ...| sed 's/^\([^\r]*\)\r\1$/\1/'|...

Handling overprinted underlining, selective bold/highlight, or merging 
overprinted lines, is out of scope or an exercise for the reader. ;^>

Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada

Problem reports:
Unsubscribe info:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]