Den sön 3 jan. 2021 kl 16:43 skrev Martin Rupp <martin.rupp_at_nefkom.net>:
> Dear Mr. Sahlberg,
>
> thank you for your answer.
>
> I have played with the client hook pre-lock script of Tortoise SVN. The
> idea is very good.
> The first argument is a temporary file. It should contain the paths to the
> files which should be locked.
> But unfortunately this file is empty.
> Therefore I am not able to add the paths to the log message via client
> hook scripts.
>
I saw this as well. This is surely a bug in TortoiseSVN which should be
reported in their mailing list (https://tortoisesvn.net/community.html).
> The svn lock command should know which file(s) in which working copy is
> (are) locked.
>
> I don't understand why this command doesn't transfer this information to
> the server and doesn't show it in the log or better in the message which is
> shown if another user wants to get a lock on this file.
>
Probably it wasn't deemed necessary for the server to know this information
when the network protocols were designed. (As you can see in JIRA one of
the developers expressed privacy concerns regarding transmitting too much
information from the client to the server). Any client is free to add this
information in the lock message and to get and display the lock message
whenever it gets the warning message W160035.
> I need much time to find the working copy (plus client computer name)
> where a file is locked because I don't know where it is locked.
>
I can agree with you, in fact I even have the same problem myself. So far I
have solved it by being strict on only using one specific WC when I edit
the files that require locking (ie, binary files).
> A software which provides the locking feature should also told the user
> where it is locked.
> I know this from other software products.
>
Subversion is an open source software and patches are always welcome (in
dev_at_subversion.apache.org).
Kind regards,
Daniel Sahlberg
Received on 2021-01-03 20:23:17 CET