"Ignore ancestry" option on reverting changes
Kazutoshi Satoda
Since merge operation with recent versions of Subversion got very slow,
a simple reverting from Log dialog, which invokes a reverse merge,
also became painfully slow.
While searching for this issue on the web, I found that
"ignore ancestry" can be used as a workaround. As implied, it doesn't
care about svn:mergeinfo. I bet this is not a problem for reverting a
change.
Then, could you please use "ignore ancestry" on reverting changes?
Here is a comparison with the CLI and TSVN repository r18441. Both
commands are run repeatedly till the time become almost stable between
each runs.
$ svn --version -q
1.6.6-SlikSvn-tag-1.6.6@40358-WIN32
$ time svn merge --dry-run -r18439:18438 .
--- Reverse-merging r18439 into '.':
U src\Utils\MiscUI\ProgressDlg.cpp
real 0m39.862s
user 0m0.015s
sys 0m0.046s
$ time svn merge --dry-run -r18439:18438 --ignore-ancestry .
--- Reverse-merging r18439 into '.':
U src\Utils\MiscUI\ProgressDlg.cpp
real 0m9.005s
user 0m0.015s
sys 0m0.000s
--
k_satoda
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=757&dsMessageId=2442187
To unsubscribe from this discussion, e-mail: [dev-uns...@tortoisesvn.tigris.org].
Stefan Küng
> Hi,
>
> Since merge operation with recent versions of Subversion got very slow,
> a simple reverting from Log dialog, which invokes a reverse merge,
> also became painfully slow.
>
> While searching for this issue on the web, I found that
> "ignore ancestry" can be used as a workaround. As implied, it doesn't
> care about svn:mergeinfo. I bet this is not a problem for reverting a
> change.
Well, it is: if you revert the changes from a revision without adjusting
the mergeinfo, that revision will still be marked as merged even though
it isn't anymore.
Stefan
--
___
oo // \\ "De Chelonian Mobile"
(_,\/ \_/ \ TortoiseSVN
\ \_/_\_/> The coolest Interface to (Sub)Version Control
/_/ \_\ http://tortoisesvn.net
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=757&dsMessageId=2442502
Kazutoshi Satoda
>> While searching for this issue on the web, I found that
>> "ignore ancestry" can be used as a workaround. As implied, it doesn't
>> care about svn:mergeinfo. I bet this is not a problem for reverting a
>> change.
>
> Well, it is: if you revert the changes from a revision without adjusting
> the mergeinfo, that revision will still be marked as merged even though
> it isn't anymore.
Really?
My test (sorry, not with TSVN repository) shows that reverting a merge
(a revision X which recorded svn:mergeinfo) does adjust svn:mergeinfo
correctly (set it back to the state before X) with "--ignore-ancestry".
This is what I assumed.
Do you know an actual case where my assumption fails? I'll try that.
--
k_satoda
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=757&dsMessageId=2442507
Kazutoshi Satoda
> Stefan Küng wrote:
>>> While searching for this issue on the web, I found that
>>> "ignore ancestry" can be used as a workaround. As implied, it doesn't
>>> care about svn:mergeinfo. I bet this is not a problem for reverting a
>>> change.
>>
>> Well, it is: if you revert the changes from a revision without adjusting
>> the mergeinfo, that revision will still be marked as merged even though
>> it isn't anymore.
Sorry, I found that my first wording was confusing. Let me rephrase.
Saying "it doesn't care about svn:mergeinfo", I meant that it
(--ignore-ancestry option) doesn't involve special cares (walking the
whole target searching for subtree svn:mergeinfo, mixed revisions, and
switched nodes, etc...) which causes the slowdown on normal merge.
--
k_satoda
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=757&dsMessageId=2442511
Stefan Küng
I'll change the code to use --ignore-ancestry for the undo functions in
the log dialog.
Stefan
------------------------------------------------------
http://tortoisesvn.tigris.org/ds/viewMessage.do?dsForumId=757&dsMessageId=2442528