[vsgrab-dev] CVSGrab - reconstructing version history
Brought to you by:
ludovicc
From: John V. <ja...@gm...> - 2005-11-06 16:47:52
|
Hi, Firstly, CVSGrab is amazing. I started off today writing a script around wget to export a project that is only available using FreeBSD-CVSweb 2.0.6, and eventually gave up. It was intoxicating to watch CVSGrab download all the most recent versions, complete with CVS directories. I have a few non-frequently asked questions. Why is the default file type binary? I needed to add the following entries in FileTypes.properties, if only to prevent the scary messages on my screen. *.js=3Dtext *.dtd=3Dtext *.php=3Dtext *.conf=3Dtext Now that I have a checkout of the project, I know that I only need the history of 4 files, and there is no branches to consider. I'm guessing that a for loop could fetch each version from 1.1 to 1.900+ ? I could then replay them against a new local cvs repository, and the changes should be accurate, with only the timestamps different from the official, but closed, tree. However, I would like the history of the entire project if possible.=20 Ideally CVSGrab could save the history as ,v files files, but I'm not fussy! Can this be achieved with 2.2.2 plus a bit of shell scripting, or is the in-memory object tree in CVSGrab better suited for performing this? If its best developed inside CVSGrab, is there any way that I can build CVSGrab without Maven, and if possible, without Ant. I know they are great tools, but my linux box wilts under the weight of blackdown Java, and I would rather have an obscure once-off shell script to build it, if at all possible. Also, the link to JRCS on the CVSGrab homepage is broken. Where is it now hosted? The only source code I can find is ironically available via ViewCVS: http://cvs.jrcs.codehaus.org/jrcs/src/java/org/apache/commons/jrcs/ -- John |