|File: [cvsweb.bsd.lv] / cvsweb / README (download)
Revision 22.214.171.124, Thu Nov 7 09:44:35 2019 UTC (3 years, 1 month ago) by schwarze
CVS Tags: cvsweb-2_1
Changes since 126.96.36.199: +2 -1 lines
Fix CVS identifiers: remove "FreeBSD" ids that were never functional
in the KNU/SCOP repository, remove "Idaemons" ids that got frozen at
some random point in the history, add "knu" ids reporting the latest revision
from the KNU/SCOP repo merged into this branch, and add "Id" identifiers
to report the current revision in the SCHWARZE repository.
$Id: README,v 188.8.131.52 2019/11/07 09:44:35 schwarze Exp $
$knu: README,v 1.2 2002/07/03 20:45:43 scop
The cgi-script cvsweb.cgi is written by Bill Fenner <firstname.lastname@example.org>
(see ) for the freebsd project. It allows browsing of
CVS-repositories (see ) with an HTML-browser.
This version is based on Bill's script and is hacked to fit our needs,
so don't blame him if anything doesn't work ..
- a colored side by side diff
- easier navigation
- mime-type aware checkout which allows for browsing
of version controlled HTML-pages
- easier configuration - everything configurable is in one
external file now
- added cvs-annotate - this requires cvs >= 1.9
(tested with cvs 1.10)
- a better way of viewing timestamps: they're shown as age
(something like 'changed 2 days 4 hours ago')
- sorting by filename / age - just click on the corresponding
- urlencodes all filenames so files with spaces and special
characters in their names are viewable as well.
- selectionbox for choosing different CVS-roots
- overview of last revisionnumber, age and logentry in the
- you can check out the current revision of a file if you are
in the directory view (idea from Nick Brachet)
- gzip'ed output if the browser accepts this; this can reduce the
size (and download time) by a factor of 4 (directory view) to
10 (colored diff; the size of the colored diff could be reduced
with CSS .. anyone ?). The use of gzip has been in the very first
version of cvsweb (by Bill Fenner), but has been removed because
it wasn't easy to determine if a browser actually understands it. Today
the browsers state if they accept the Content-encoding gzip..
.. but MSIE lies about this (so cvsweb disables gzip-encoding in this
- Tag based browsing. An easy and what I think intuitive way to browse a
CVS tree based on tags, with filtering of files and CVS logs to only
show whats relevant for the selected tag.
- "Pretty-printed" checkouts for text based files. My intention is to
eventually add cross-referencing to this, but for now the file is only
converted to HTML with automatic detection of various links.
- "Click-on-version" diff selection.
- sorting for every column now
- uses 'cvs co' for checkout instead of rcs co
- Ability to show directory descriptions from CVSROOT/descriptions
cvsweb.cgi became quite huge now and is bad style code (I am not a
perl-programmer - this is what I do in my spare time!) so this
should be rewritten.
I'd propose a Java-servlet which could use the C/S-protocol .. anyone ?
And what about using mod_perl, and rewriting it into a set of proper Perl-5
This software is distributed under the terms of the BSD-licence.
You can download the latest version here:
$ cvs -d email@example.com:/cvs co cvsweb
 for the original Version of cvsweb.cgi, see
 for cvs, see <http://www.nongnu.org/cvs/>