[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Searching private SVN Repository - how? best ways?

From: Michael B. Parker <MBParker_at_Cytex.com>
Date: 2006-06-18 13:51:48 CEST

I'm looking at creating a wiki (plus general repository) with
Subversion. And subversion seems cool, but how do you search it?
Especially if the repository is private or semi-private? This is all
I can find (from least to most powerful roughly):

1. http://subversion.tigris.org/tools_contrib.html only seems to
list wcgrep <http://subversion.tigris.org/tools_contrib.html#wcgrep>
which only searches the working copy and (the killer) is brute-force (no
indexing).
2. One could make a local copy of the whole repository (or the
parts one has access to), or net mount a working copy (as this user does
<http://svn.haxx.se/users/archive-2006-06/0083.shtml> ), and then use a
local desktop indexer (as http://Toolbar.MSN.com
<http://toolbar.msn.com/> ); but making a copy of the whole repository
(if each user does it) is inefficient (if not impractical) if the
repository is large, plus a security risk (full copies floating around)
if the data is private; only way to do this properly would be to run the
indexer on the server and come up with some web interface.
3. An "add search to your website" service, as www.FreeFind.com
<http://www.freefind.com/> will do it: see How to Index Password
Protected Pages -- FreeFind.com
<http://www.freefind.com/library/howto/passwords/> , though it doesn't
have knowledge of source code and histories.
4. Google.com <http://www.google.com/> and other major public
search engines would be great; though they still don't address the
special issues of source control (versioning, symbol definition & cross
referencing), being used by everyone, this would still be very desirable
as it brings visitors to the site. Except that it seems they still
can't search private content (at least easily) without compromising the
security. I've posted a question asking for fixes (Have public search
engines index private content (but still require users sign it to get)
--How?
<http://groups.google.com/group/google-sitemaps/browse_frm/thread/832c46
b66108949f/82d6b8f33e806833?lnk=st&q=Have+public+search+engines+index+pr
ivate+content&rnum=1#82d6b8f33e806833> ) but haven't found any good
solutions yet; still, a number of sites as www.Experts-Exchange.com do
this trick
5. www.ViewVC.Org <http://www.viewvc.org/> talks of "Bonsai
<http://www.mozilla.org/projects/bonsai/> -like repository query
facilities" and "Bonsai <http://www.mozilla.org/projects/bonsai/> -like
repository query facilities" and "MySQL <http://www.mysql.com/> -
Needed to use the commit database query functionality", but I can't find
any info? Does that search it? Does it index it? (this post
<http://svn.haxx.se/users/archive-2006-06/0083.shtml> suggests it does,
but how well?)
6. this post <http://svn.haxx.se/users/archive-2006-06/0084.shtml>
ref'd me to www.opensolaris.org/os/project/opengrok
<http://www.opensolaris.org/os/project/opengrok> which does it:
"typical case where OpenGrok is deployed on a webserver for an active
source repository, which is updated automatically using scheduled cron
jobs. SRC_ROOT refers to the directory containing your source tree.
OpenGrok analyzes the source tree and builds a search index along with
cross-referenced hypertext versions of the source files." (says here
<http://www.opensolaris.org/os/project/opengrok/manual/> ); the main
page has a "Comparison with Similar Tools" which suggests it beats
ViewVC.Org on every feature.
7. www.cenqua.com/fisheye <http://www.cenqua.com/fisheye> (a sort
of "ViewCVS on steroids") absolutely does it (and much more), but that
costs $1500/server (free for open source code, but then again that's not
private).

 

Thoughts? Preferences? Thanks!

 

-Mike Parker, MIT CS grad (here helping build www.CommuniDB.com
<http://www.communidb.com/> , an XML wiki of sorts likely based on
Subversion)

 

 

Aside: ViewVC plus many other subversion tools are missing from
http://subversion.tigris.org/tools_contrib.html; in contrast,
freshmeat.net Browse project tree - Topic Software Development Version
Control Subversion <http://freshmeat.net/browse/1109/> has several;
might want to have http://subversion.tigris.org
<http://subversion.tigris.org/> link to lists such as this one..
Received on Sun Jun 18 13:55:48 2006

This is an archived mail posted to the Subversion Users mailing list.