DIFF - Strange Behaviour of Intertrac Diff for a Decentral Branch

0 views
Skip to first unread message

Ilias Lazaridis

unread,
Nov 11, 2006, 8:00:39 AM11/11/06
to Trac Development
based on the suggestion of cboos to use the diff mechanism of trac:

CODE - Flexible TicketQuery Macro with extended Query functionality
http://groups.google.com/group/trac-dev/msg/d8255aff2429608d

i've provided a difference report of my decentral development branch in
relation to the original trac.

In order to achieve this, I needed to placed 2 copies of the trac
source-tree within my personal svn repository (decentral in relation to
trac, e.g. other server).

One source is the original one within /sync/trac.

The other is the branch i'm actually working on within /infra/trac-dev.

Based on those I can create conveniently a full difference report:

http://dev.lazaridis.com/base/intertrac/diff:sync/trac@head:infra/trac-dev@head

-

The Problem:

I am not exactly sure if this is a bug, but there are 2 issues:

1.) binaries are displayed as changed, although I've not changed them:

e.g.:
# infra/trac-dev/htdocs/newticket.png (modified) (previous)
# infra/trac-dev/htdocs/extlink.gif (modified) (previous)

2.) for some reason, the nam "Borgstr..." becomes disrupted

4 # Copyright (C) 2003-2005 Jonas Borgström <jo...@edgewall.com>
4 # Copyright (C) 2003-2005 Jonas Borgstrm <jo...@edgewall.com>

the funny thing is, that both locations do not state the name
correctly:

http://dev.lazaridis.com/base/browser/sync/trac/trac/ticket/api.py#L4

http://dev.lazaridis.com/base/browser/infra/trac-dev/trac/ticket/api.py#L4

-

My server is a Suse 9.3, my local environment windows (command-line
svn and tortoise gui client).

I've simply checked out the trac repo, exported it to a location and
commited it to my working branch.

.

--
http://dev.lazaridis.com/base/wiki/PlanDecentralBranching

Christian Boos

unread,
Nov 11, 2006, 9:33:12 AM11/11/06
to trac...@googlegroups.com
Ilias Lazaridis wrote:
> ...

> http://dev.lazaridis.com/base/intertrac/diff:sync/trac@head:infra/trac-dev@head
>
> -
>
> The Problem:
>
> I am not exactly sure if this is a bug, but there are 2 issues:
>
> 1.) binaries are displayed as changed, although I've not changed them:
>
> e.g.:
> # infra/trac-dev/htdocs/newticket.png (modified) (previous)
> # infra/trac-dev/htdocs/extlink.gif (modified) (previous)
>

This is apparently a bug, though I'm not sure if it's a Subversion one
(svn_repos_dir_delta) or ours. I'll check.

> 2.) for some reason, the nam "Borgstr..." becomes disrupted

> ...
> http://dev.lazaridis.com/base/browser/sync/trac/trac/ticket/api.py#L4
>

That one looks correct, byte-wise. It's only that the encoding is not
UTF set to or auto-detected to UTF-8 . See
http://trac.edgewall.org/ticket/3689.

> http://dev.lazaridis.com/base/browser/infra/trac-dev/trac/ticket/api.py#L4
>

That looks like an import issue.
Can you get to the corresponding bytes for that line? (and the ones from
the other file, for comparison).

The "canonical" byte-content should be:

$ head.exe -4 trac/ticket/api.py | tail.exe -1 | od.exe -c
0000000 # C o p y r i g h t ( C )
0000020 2 0 0 3 - 2 0 0 5 J o n a s
0000040 B o r g s t r 303 266 m < j o n a
0000060 s @ e d g e w a l l . c o m > \r
0000100 \n
0000101

(with head/tail/od from the cygwin tools)

-- Christian

Ilias Lazaridis

unread,
Nov 12, 2006, 5:19:27 AM11/12/06
to Trac Development
Christian Boos wrote:
> Ilias Lazaridis wrote:
> > ...
> > http://dev.lazaridis.com/base/intertrac/diff:sync/trac@head:infra/trac-dev@head
> >
> > The Problem:
> >
> > I am not exactly sure if this is a bug, but there are 2 issues:
> >
> > 1.) binaries are displayed as changed, although I've not changed them:
> >
> > e.g.:
> > # infra/trac-dev/htdocs/newticket.png (modified) (previous)
> > # infra/trac-dev/htdocs/extlink.gif (modified) (previous)
>
> This is apparently a bug, though I'm not sure if it's a Subversion one
> (svn_repos_dir_delta) or ours. I'll check.

ok

> > 2.) for some reason, the nam "Borgstr..." becomes disrupted
> > ...
> > http://dev.lazaridis.com/base/browser/sync/trac/trac/ticket/api.py#L4
>
> That one looks correct, byte-wise. It's only that the encoding is not
> UTF set to or auto-detected to UTF-8 . See
> http://trac.edgewall.org/ticket/3689.

I don't understand.

What steps do I have to take?

> > http://dev.lazaridis.com/base/browser/infra/trac-dev/trac/ticket/api.py#L4
>
> That looks like an import issue.
> Can you get to the corresponding bytes for that line? (and the ones from
> the other file, for comparison).

local copy form
http://dev.lazaridis.com/base/browser/sync/trac/trac/ticket/api.py

Jonas Borgstrm <jo...@edgewall.com>
4A 6F 6E 61 73 20 42 6F 72 67 73 74 72 6D 20 3C 6A 6F 6E 61 73 40 65 64
67 65 77 61 6C 6C 2E 63 6F 6D 3E

local copy from
http://dev.lazaridis.com/base/browser/infra/trac-dev/trac/ticket/api.py

Jonas BorgstrΓ¶m <jo...@edgewall.com>
4A 6F 6E 61 73 20 42 6F 72 67 73 74 72 C3 B6 6D 20 3C 6A 6F 6E 61 73 40
65 64 67 65 77 61 6C 6C 2E 63 6F 6D 3E

Ilias Lazaridis

unread,
Nov 13, 2006, 1:37:38 PM11/13/06
to Trac Development

Christian Boos

unread,
Nov 14, 2006, 7:25:01 AM11/14/06
to trac...@googlegroups.com
Ilias Lazaridis wrote:
> Christian Boos wrote:
>
>> Ilias Lazaridis wrote:
>>
>>> ...
>>> http://dev.lazaridis.com/base/intertrac/diff:sync/trac@head:infra/trac-dev@head
>>>
> ...
>
> permanent link:
>
> http://dev.lazaridis.com/base/intertrac/diff:sync/trac@110:infra/trac-dev@110
>
>

Ok, thanks, still didn't get time to look at this, but will. The
corresponding revision number is then helpful, but I noticed that this
happens also on the query changes URL you gave elsewhere...

> btw: would it be much effort to implement something like this:
>

Yes ;)

I'm not sure it's worth going into the details, it would require to get
svn informations from a remote source (#493), fetch all the content to
be able to diff them, etc. So even if this would be theoretically
possible, I'm not sure it's the right approach. Better use a distributed
VCS, svk in this case. I've read that at the SVN summit, they talked
about adding distributed capabilities to SVN in future releases. Anyone
has more details about this?

-- Christian

Ilias Lazaridis

unread,
Nov 14, 2006, 9:01:02 AM11/14/06
to Trac Development
Christian Boos wrote:
> Ilias Lazaridis wrote:
...

> > btw: would it be much effort to implement something like this:

> I'm not sure it's worth going into the details, it would require to get
> svn informations from a remote source (#493), fetch all the content to
> be able to diff them, etc. So even if this would be theoretically
> possible, I'm not sure it's the right approach. Better use a distributed
> VCS, svk in this case.

would trac support this "decentral-diff-report" if I used svk instead
of svn?

> I've read that at the SVN summit, they talked
> about adding distributed capabilities to SVN in future releases. Anyone
> has more details about this?

.

--
http://dev.lazaridis.com/base/wiki/PlanDecentralBranching

Christian Boos

unread,
Nov 14, 2006, 11:07:25 AM11/14/06
to trac...@googlegroups.com
Ilias Lazaridis wrote:
> Christian Boos wrote:
>
>> ... Better use a distributed VCS, svk in this case.

>>
>
> would trac support this "decentral-diff-report" if I used svk instead
> of svn?
>

Yes, as you would have the "vendor" branch mirrored in your local
repository, so it's really only branch-to-branch diff.

I've found a nice picture of the SVK workflow here:

http://www.bieberlabs.com/wordpress/archives/2005/03/19/explaining-the-svk-workflow/

but it's a bit outdated.

More recent docs are on the same site:
http://www.bieberlabs.com/wordpress/svk-tutorials/

and of course, there's the official SVK site (but it's a bit hard to
find clear and concise documentation there):
http://svk.elixus.org/view/HomePage

Finally, this might summarize the whole thing in an understandable way ;)
http://pickscrape.woobling.org/svk-visual-guide.pdf

But I admit I haven't used SVK in a while, being a bit frustrated by the
lack of responsiveness of its command line on Windows (on Linux, there
are tricks to speed this up).

-- Christian


Reply all
Reply to author
Forward
0 new messages