You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Lorenz <lo...@yahoo.com> on 2015/05/18 15:54:03 UTC
getting desperate because of assumed regression after 1.8.11 (was: Can someone please verfiy this "svn -v st" behaviour with file externals?)
Hi all,
I'm getting desperate because I think I've spotted a bug/regression in
svn since 1.8.11 and in spite of several post about the topic (both to
users and dev) noone seem interested in verifying or rejecting my
report.
>can someone please check if this realy is a regression, or if it is
>just me doing something wrong.
>
>I have a problem with "svn -v status" showing an unexpected result for
>file externals.
>
>This seems to be a regression that happended with svn 1.8.12/13
>
>I've tried this with the binaries bundled with TSVN 1.8.11 (svn
>1.8.13) and TSVN trunk nightly (build against
>svn.apache.org/repos/asf/subversion/branches/1.9.x)
>
>I've also tried with the 1.8.13 windows builds from
>http://sourceforge.net/projects/win32svn/ and
>http://www.visualsvn.com/downloads/ with the same results.
>
>
>Using svn 1.8.11 binaries (bundled with TSVN 1.8.10) the is effect is
>not reproducible.
>
>
>After freshly checking out a working copy (containing file externals)
>at first everything is ok.
>
>C:\tmp\svn-file-ext-tests>svn -v status wc
> 2 2 me wc
> X 2 1 me wc\empty-main.txt
> 2 2 me wc\sub
> X 2 1 me wc\sub\empty-sub.txt
>
>
>But after a "svn up" I get something like the following:
>
>C:\tmp\svn-file-ext-tests>svn -v st wc
> 2 2 me wc
> X 2 ? ? wc\empty-main.txt
> 2 2 me wc\sub
> X 2 ? ? wc\sub\empty-sub.txt
>
>
>PS: After producing the the error with a 1.8.13 version on a working
>copy, switching back to 1.8.11 also shows the errror on this working
>copy.
>
>
>>>>>> reproduction receipt (windows batch)
>>>
>
>svnadmin create <repo path>
>
>mkdir wc\main\sub
>echo empty >wc\empty.txt
>
>svn import --username me -m "" wc <repo path>
>svn checkout --force <repo path> wc
>
>svn propset svn:externals "^/empty.txt empty-main.txt" wc
>svn propset svn:externals "^/empty.txt empty-sub.txt" wc/sub
>svn commit --username me -m "" wc
>
>svn checkout <repo path> wc2
>svn -v status wc2
>
>svn update wc2
>svn -v st wc2
>
><<
><<<<<
--
Lorenz
Re: getting desperate because of assumed regression after 1.8.11
Posted by Philip Martin <ph...@wandisco.com>.
Stefan Sperling <st...@elego.de> writes:
> On Mon, May 18, 2015 at 01:54:03PM +0000, Lorenz wrote:
>> Hi all,
>>
>> I'm getting desperate because I think I've spotted a bug/regression in
>> svn since 1.8.11 and in spite of several post about the topic (both to
>> users and dev) noone seem interested in verifying or rejecting my
>> report.
>
> I can reproduce based on your script. Please file an issue.
> Thanks for your bug report, and for being so patient, and for not giving up :)
After a quick scan of the log for likely candidates the culprit appears
to be:
------------------------------------------------------------------------
r1660023 | svn-role | 2015-02-16 04:01:29 +0000 (Mon, 16 Feb 2015) | 13 lines
Merge the 1.8.x-issue4411 branch:
* r1513479, r1625522
Fix issue #4411, 'parent-relative file external can corrupt working copy'.
Justification:
Broken working copies are bad.
Notes:
This also adds a test case for issue #4519 which was fixed in 1.8.11.
Branch:
^/subversion/branches/1.8.x-issue4411
Votes:
+1: stsp, rhuijben, stefan2
------------------------------------------------------------------------
With r1660022:
At revision 2.
2 2 pm wc
2 1 pm wc/e
X 2 1 pm wc/em
2 2 pm wc/s
X 2 1 pm wc/s/es
With r1660023:
At revision 2.
2 2 pm wc
2 1 pm wc/e
X 2 ? ? wc/em
2 2 pm wc/s
X 2 ? ? wc/s/es
--
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Re: getting desperate because of assumed regression after 1.8.11
(was: Can someone please verfiy this "svn -v st" behaviour with file
externals?)
Posted by Stefan Sperling <st...@elego.de>.
On Mon, May 18, 2015 at 01:54:03PM +0000, Lorenz wrote:
> Hi all,
>
> I'm getting desperate because I think I've spotted a bug/regression in
> svn since 1.8.11 and in spite of several post about the topic (both to
> users and dev) noone seem interested in verifying or rejecting my
> report.
I can reproduce based on your script. Please file an issue.
Thanks for your bug report, and for being so patient, and for not giving up :)
I used this command sequence with 'svn' built from trunk:
[[[
svnadmin create /tmp/repos
mkdir -p /tmp/wcwmain/sub
echo empty > /tmp/wc/empty.txt
svn import -m "" wc file:///tmp/repos
svn checkout --force file:///tmp/repos /tmp/wc
svn propset svn:externals "^/empty.txt empty-main.txt" /tmp/wc
svn propset svn:externals "^/empty.txt empty-sub.txt" /tmp/wc/sub
svn commit -m "" /tmp/wc
svn checkout file:///tmp/repos /tmp/wc2
svn -v status /tmp/wc2
svn update /tmp/wc2
svn -v st /tmp/wc2
]]]
+ svn -v status /tmp/wc2
2 2 stsp /tmp/wc2
X 2 1 stsp /tmp/wc2/empty-main.txt
2 1 stsp /tmp/wc2/empty.txt
+ svn update /tmp/wc2
Updating 'wc2':
Fetching external item into 'wc2/empty-main.txt':
External at revision 2.
At revision 2.
+ svn -v st /tmp/wc2
2 2 stsp /tmp/wc2
X 2 ? ? /tmp/wc2/empty-main.txt
2 1 stsp /tmp/wc2/empty.txt
Re: getting desperate because of assumed regression after 1.8.11 (was: Can someone please verfiy this "svn -v st" behaviour with file externals?)
Posted by Lorenz <lo...@yahoo.com>.
Lorenz wrote:
>[...]
>>I have a problem with "svn -v status" showing an unexpected result for
>>file externals.
>>
>>This seems to be a regression that happended with svn 1.8.12/13
>[...]
this ended up as Issue 4580, got fixed and ported to 1.8.x and 1.9.x.
Thanks guys
I've run tests with a nightly build of TSVN (build against the 1.9.x
branch) from last night, and can confirm that file externals are
working again.
--
Lorenz