Re: Bulk-Add

[schmolle]

Hi all,

A not-so-very-nice, but quite effective way of doing it:

$ cd $WORKINGDIR
$ find . \( -type d -name "CVS" -prune \) -o \( -type d -exec cvs add {} \; \) 
2>/dev/null
$ find . \( -type d -name "CVS" -prune \) -o \( -type f -exec cvs add {} \; \) 
2>/dev/null

The first find will add all your directories; error output goes to /dev/null. 
The second one takes care of all your files.

Files / directories that already exist in the repository will just produce a 
harmless error.

The first part of the find statement avoids going down the CVS directories in 
your workspace. (cvs does not like it when you try to 'add' those)

Disadvantage: this is rather time-consuming, because a seperate cvs operation 
is done for every directory or file.

Advantage: This will (iirc) handle file names with spaces, if you really must 
have those. It is also a big hammer to crack all sizes of nuts. ymmv.

If you are the only person using a repository (or one of a very few), you might 
want to try parallelising the effort; something like:

#!/bin/sh
USER=MyName
WORKINGDIR=/home/${USER}/MyWorkingDir
TMPFILE=/home/${USER}/CVSTEMPFILE

# Find all dirs and add them
find . \( -type d -name "CVS" -prune \) -o \( -type d -print \) | while read 
DIR; do
        cvs add ${DIR} 2>/dev/null &
done 
# Wait until all are done
wait

find . \( -type d -name "CVS" -prune \) -o \( -type f -print \) | while read 
FILE; do
        cvs add ${FILE} 2>/dev/null &
done
wait

# EOF

Try the difference between doing this and doing the same without the 
ampersands. It is possible there are huge speed improvements It may also be 
possible you find yourself locking yourself out until the cows come home. I 
have had very good results doing this, but ony after I found a good balance 
between dropping everything in the background and doing that more selectively.


> A recursive bulk cvs add is more work

Only for the machine. :) But indeed, the above approach runs once---done. Using 
CVS's interpretation of what is and isn't known in your workspace can be more 
than a bit cumbersome.

>Probably easier just to run the update script a bunch of times

With lots of subdirs of subdirs, the large-hammer approach could very well be 
faster.

TTYL,

Schmolle


ps: I'm writing this code of the top of my head. TEST IT BEFORE YOU USE IT! And 
read the find man page. I'll see if I can verify my claims today and post 
additional comments if need be.