try this script..
#!/bin/bash
NULL="_"
for f in `svn st|grep -v ^\?|sed s/.\ *//`;
do LIST="${LIST} $f $NULL on";
done
dialog --checklist "Select files to commit" 30 60 30 $LIST 2>/tmp/svnlist.txt
svn ci `cat /tmp/svnlist.txt|sed 's/"//g'`
You basically put the files you want to commit on the command line
svn ci file1 file2 dir1/file3
Due to my subversion state, I had to get creative. svn st showed M,A and ~ statuses. I only wanted M and A so...
svn st | grep ^[A\|M] | cut -d' ' -f8- > targets.txt
This command says find all the lines output by svn st that start with M or A, cut using space delimiter, then get colums 8 to the end. Dump that into targets.txt and overwrite.
Then modify targets.txt to prune the file list further. Then run below to commit:
svn ci -m "My commit message" --targets targets.txt
Probably not the most common use case, but hopefully it helps someone.
I make a (sub)folder named "hide", move the file I don't want committed to there. Then do my commit, ignoring complaint about the missing file. Then move the hidden file from hide back to ./
I know of no downside to this tactic.
Use changelists. The advantage over specifying files is that you can visualize and confirm everything you wanted is actually included before you commit.
$ svn changelist fix-issue-237 foo.c
Path 'foo.c' is now a member of changelist 'fix-issue-237'.
That done, svn now keeps things separate for you. This helps when you're juggling multiple changes
$ svn status
A bar.c
A baz.c
--- Changelist 'fix-issue-237':
A foo.c
Finally, tell it to commit what you wanted changed.
$ svn commit --changelist fix-issue-237 -m "Issue 237"
Besides listing the files explicitly as shown by unwind and Wienczny, you can setup change lists and checkin these. These allow you to manage disjunct sets of changes to the same working copy.
You can read about them in the online version of the excellent SVN book.
Source: Stackoverflow.com