>If your script really has 'echo off' and not '@echo off' then your script
>produces output. And you should check the documentation on how that is handled
>for the specific hooks. For pre-lock this has side effects:
>[[
># If the hook program outputs anything on stdout, the output string will # be
>used as the lock token for this lock operation. If you choose to use # this
>feature, you must guarantee the tokens generated are unique across # the
>repository each >time.
>]]
Sorry, my mistake, you are correct. I corrected it to " @echo off".
I tested it 5 minutes ago, same result for either "@echo off" or "echo off":
- during commit all 3 hooks are executed, commit is successfull
- during lock no hook is executed, lock failed
> What operation did you perform to test the lock hook?
> svn lock <something>?
> The lock script is invoked as part of committing.
Yes. I used testing SVN instance on my machine, so SVN bin is not on the PATH.
Copied/pasted:
<code>
C:\work\dzcvk82\svn_views\test6\test001>dir
Výpis adresáře C:\work\dzcvk82\svn_views\test6\test001
27.11.2013 15:05 <DIR> .
27.11.2013 15:05 <DIR> ..
27.11.2013 15:05 135 testfile.txt
Souborů: 1, Bajtů: 135
Adresářů: 2, Volných bajtů: 6 574 776 320
C:\work\dzcvk82\svn_views\test6\test001>c:\software\Subversion\bin\svn.exe lock
testfile.txt
svn: E200035: sqlite[S19]: LOCK.lock_token may not be NULL
svn: E200035: Additional errors:
svn: E200035: sqlite[S19]: LOCK.lock_token may not be NULL
C:\work\dzcvk82\svn_views\test6\test001>
</code>
Perhaps the the lock fails before the hook execution.
Vena Kovar
