Glad that 'changed' gives you what you want, up to parsing the output.
If you need more control, you can write scripts that use the svn_repos_*
API's directly (via the Perl bindings).
And, for next time, please try to get some feedback on your ideas for
patches before you start coding them. It will save your time and ours.
Thanks for your contribution,
Daniel
John Huddleston wrote on Sun, 8 Jun 2008 at 09:28 -0600:
> Daniel
>
> Thank you for your quick reply. The changed option does indeed give the
> correct content; however, it has the 'U" in the front of the line and the
> svnlog.pl that I have created returns the full URL to the user with the http
> prepended in the front of the line. Thus, the 'tree' option with the
> "--full-paths" option is preferred.
>
> I have a modified version of the original code I sent out that checks to see
> if the user requested both the 'tree' and the '-r' option. This could be
> submitted as a "PATCH" item.
>
> We can close this email thread since it appears that my request is an
> enhancement and not a bug fix.
>
> John Huddleston
>
> -----Original Message-----
> From: Daniel Shahaf [mailto:d.s_at_daniel.shahaf.co.il]
> Sent: Sunday, June 08, 2008 7:23 AM
> To: John Huddleston
> Cc: dev_at_subversion.tigris.org; users_at_subversion.tigris.org
> Subject: RE: SVNLOOK tree strictly obeying the version
>
> [ CCing users@ list, since it's no longer about a patch ]
>
> John Huddleston wrote on Sun, 8 Jun 2008 at 06:46 -0600:
> > Daniel,
> >
> > I'm sorry if my original email to this group was terse.
> >
> (I was going to say the same about my last email.)
> > If you elect to use the "-r" option of the svnlook command, one would hope
> > that you would only get records that that been touched (like an existing
> > directory), added, modified, or even deleted.
> >
>
> Please try 'svnlook changed -r'.
>
> > The behavior is that there are revisions being displayed that are not the
> > ones the user requested. This is the spurious data.
> >
> > You said "It's a few megabytes of output, what's wrong with it?"
> >
> > If you do not see a problem in getting information that does not conform
> to
> > the user's request, then that must have been the original design of the
> > program.
> >
>
> Yes.
>
> 'svnlook changed -r' prints the paths that were modified in a given
> revision.
>
> 'svnlook tree -r' prints the entire versioned filesystem tree (like 'svn
> ls -R -r') as it existed in a given revision.
>
> Daniel
>
> > John Huddleston
> >
> >
> >
> > -----Original Message-----
> > From: Daniel Shahaf [mailto:d.s_at_daniel.shahaf.co.il]
> > Sent: Saturday, June 07, 2008 9:32 AM
> > To: John Huddleston
> > Cc: dev_at_subversion.tigris.org
> > Subject: RE: SVNLOOK tree strictly obeying the version
> >
> > John Huddleston wrote on Sat, 7 Jun 2008 at 09:10 -0600:
> > > Daniel
> > >
> > > I have five or sixe different subversion repositories that manifest this
> > > behavior.
> > >
> >
> > *Which* behaviour? You still have not explained why you think the
> > output is wrong, which was my original question.
> >
> > > svnlook tree --show-ids --full-paths /my/repos/path
> > >
> > > Some were created with the 1.4.5 version and some older one were created
> > > with 1.4.3
> > >
> > > Try it for your self on the subversion repository.
> > >
> > > svnlook tree --show-ids --full-paths /your/path/to/collabnet/subversion
> |
> > > grep -v r31623
> > >
> >
> > It's a few megabytes of output, what's wrong with it?
> >
> > % svnlook help tree
> > tree: usage: svnlook tree REPOS_PATH [PATH_IN_REPOS]
> >
> > Print the tree, starting at PATH_IN_REPOS (if supplied, at the root
> > of the tree otherwise), optionally showing node revision ids.
> >
> > Originally you said that that the "spurious data" was output when you
> > passed 'the "-r" option to the "svnlook tree" command'. The above
> > command does use that option, and does cannot logically trigger the bug
> > you originally reported (and tried to fix).
> >
> > (see below for more...)
> >
> > > John Huddleston
> > >
> > > -----Original Message-----
> > > From: Daniel Shahaf [mailto:d.s_at_daniel.shahaf.co.il]
> > > Sent: Sunday, June 01, 2008 10:41 AM
> > > To: John Huddleston
> > > Cc: dev_at_subversion.tigris.org
> > > Subject: Re: SVNLOOK tree strictly obeying the version
> > >
> > > John Huddleston wrote on Sat, 31 May 2008 at 06:59 -0600:
> > > > Hi,
> > > >
> > > >
> > > > I was setting up a post-commit svnlog.pl perl script for a customer
> and
> > > they
> > > > wanted the viewvc URL of the changed files sent to them via email.
> >
> > % svnlook help changed
> > changed: usage: svnlook changed REPOS_PATH
> >
> > Print the paths that were changed.
> >
> > Daniel
> >
> > > > During
> > > > the process of programming I noticed that the "-r" option to the
> > "svnlook
> > > > tree" command returned some spurious data.
> > > >
> > >
> > > Can you provide a recipe (starting with an empty repository) that causes
>
> > > svnlook to produce spurious data? Or at least, explain to us what the
> bug
> >
> > > is, before you start fixing it :)
> > >
> > > By "recipe", I mean a script of the form:
> > >
> > > svnadmin create repos
> > > svn co `pwd`/repos wc
> > > cd wc
> > > ### make some commits
> > > svnlook tree ../repos -r REV
> > >
> > > Thanks,
> > >
> > > Daniel
> > >
> > > P.S. For reference, please read
> > > http://subversion.tigris.org/hacking.html#patches. Following the
> > > guidelines there makes it easier for us to process your patch, which
> makes
> >
> > > it more likely that it will be applied.
> > >
> > > >
> > > >
> > > > I checked out the most recent code from your subversion server and
> > edited
> > > > the main.c under
> > http://svn.collab.net/repos/svn/trunk/subversion/svnlook
> > > to
> > > > print the requested revisions.
> > > >
> > > >
> > > >
> > > > I apologize if it is not elegant but this was my first foray into
> > > modifying
> > > > any of this code. In short, I needed to pass the revision from the
> > > > svnlook_ctxt_t and then compare it to the revision in the database.
> > > >
> > > >
> > > >
> > > > I've attached an svn diff file with the changes. The work was done on
> > > > Windows working under Cygwin.
> > > >
> > > >
> > > >
> > > > John Huddleston
> > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
> > > For additional commands, e-mail: dev-help_at_subversion.tigris.org
> > >
> > >
> > >
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
> > For additional commands, e-mail: dev-help_at_subversion.tigris.org
> >
>
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: users-help_at_subversion.tigris.org
Received on 2008-06-08 17:48:40 CEST