On 2023-06-20 13:45, Daniel Sahlberg wrote:
Den tis 20 juni 2023 kl 13:00 skrev Osipov, Michael (SMD IT IN)
<michael.osi...@siemens.com <mailto:michael.osi...@siemens.com>>:
From a user's perspective this does not make sense at all:
> osipovmi@deblndw011x:~/var/Projekte/Playground
> $ echo foo > m...@apache.org <mailto:m...@apache.org>
> osipovmi@deblndw011x:~/var/Projekte/Playground
> $ LC_ALL=C svn add m...@apache.org <mailto:m...@apache.org>
> svn: E200009: 'm...@apache.org <mailto:m...@apache.org>': a peg
revision is not allowed here
I would really expect that the peg rev would used where it makes sense
and not on local paths at all.
I don't say you are wrong but in the previous thread, three persons who
have been involved in the project a long time argued that:
- Yes, it is a pita and sometimes it doesn't even make sense
- There was maybe a reason long time ago that it was thought to be more
consistent to always parse the PEG revision on all paths (however, as I
found out there is at least one case where the peg revision is not NOT
parsed). (I didn't dig through the version history and the mailing list
archives from where this was added, it is left as an exersice for the
reader).
- Changing things now would have backwards compatibility implications,
any user of the command line client (for example for the Maven SVM
Subversion Provider) would need to be aware of when this changed and
behave accordingly.
We CAN change the documentation, which is what I started looking at.
Agree, doc update is imperative here. As well as the breaking of the
behavior if this is fixed. I guess this would qualify for 2.0.0, not
before as it might break users and automated code.
M
