Formatting svn status output

[Real World]

I'm wanting to be able to write a text file with a list of all the files that have been updated since the last svn update with each file on a separate line.
To do this I need to format the output of this:

root@server # svn status -u
* 9 admin.php
* 9 index.php
* 9 functions/myfuncs.php
N 9 functions/newfile.php
Status against revision: 10

into this

root@server # svn status -u
admin.php
index.php
functions/myfuncs.php
functions/newfile.php

I am really inexperienced in the voodoo of regex but I suspect thats what is needed here. Anyone care to share their wisdom with this noob? Thanks

[Cabhan]

You don't need a regular expression at all. Instead, you can use the cut utility.

Basically, you tell cut what character separates your columns (in this instance, a space), and what columns you want (in this case, the third), and it will only give you those columns back.

Check the man page for more info.

[Real World]

Sorry I shouldn't have used quote tags there

root@server # svn status -u
       *        9   admin.php
       *        9   index.php
       *        9   functions/myfuncs.php
       N        9   functions/newfile.php
Status against revision:     10

There is more than one space between those columns. I don't think theyre tab characters either. Would cut still work?

[Cabhan]

Ah, right you are sir. Easy enough! We will resort to sed and regular expressions, then.

Basically, we want to remove everything except for the last column:

alex@danu ~/test/bash $ cat svn_output 
       *        9   admin.php
       *        9   index.php
       *        9   functions/myfuncs.php
       N        9   functions/newfile.php
alex@danu ~/test/bash $ sed -e 's/.\+ //' svn_output 
admin.php
index.php
functions/myfuncs.php
functions/newfile.php

Why does this work? I have told sed to replace 1 or more of any character, followed by a space, with nothing. Most regular expressions are "greedy", which means that they try and match as much as they possibly can. Therefore, that last space comes as late as possible, and it catches on the space just before the last column, so everything up to and including that is removed.

Make sense?

[Franklin52]

Or simply with awk, print field 3 if the line doesn't begin with Status:

awk '!/^Status/{print $3}' svn_output

Regards

[Franklin52]

Ah, right you are sir. Easy enough! We will resort to sed and regular expressions, then.

Basically, we want to remove everything except for the last column:

alex@danu ~/test/bash $ cat svn_output 
       *        9   admin.php
       *        9   index.php
       *        9   functions/myfuncs.php
       N        9   functions/newfile.php
alex@danu ~/test/bash $ sed -e 's/.\+ //' svn_output 
admin.php
index.php
functions/myfuncs.php
functions/newfile.php

Why does this work? I have told sed to replace 1 or more of any character, followed by a space, with nothing. Most regular expressions are "greedy", which means that they try and match as much as they possibly can. Therefore, that last space comes as late as possible, and it catches on the space just before the last column, so everything up to and including that is removed.

Make sense?

Cabhan,

You forgot the last line of his svn output.

Regards

[Real World]

Cabhan,

You forgot the last line of his svn output.

Regards

Indeed he did which leads to a trailing integer at the end of the file and Franklin yours didn't work perfectly either but thats my fault as again the status output was wrong :S

       *            install.php
       *        9   admin.php
       *        9   .
Status against revision:     11

install.php is a new file. I thought it displayed N instead of a * and the version number but apparently not. Sorry.
This is good stuff though I'm learning some tricks I didn't know. I appreciate it

[Franklin52]

Did you solve the problem? If not post the exact svn output and the desired output.

Regards

[Cabhan]

Sorry for forgetting that last line. In any event, I think that my approach is more easily adaptable:

alex@danu ~/test/bash $ cat svn_output 
       *            install.php
       *        9   admin.php
       *        9   .
Status against revision:     11
alex@danu ~/test/bash $ sed -e '/^Status/ d; /^Status/ ! s/.* //' svn_output 
install.php
admin.php
.

My sed expression says that for any line that starts with "Status", ignore it. Otherwise, delete everything up through the last space.

Note that my approach will _not_ work if any of the filenames have a space in them.

[ghostdog74]

if you look at the help of svn, using

 svn help status

you can see :

The out-of-date information appears in the eighth column (with -u):
'*' a newer revision exists on the server
' ' the working copy is up to date

you could possible try to substring from nineth character onwards.