Could we partially delete? (As well as add the large readme.txt too obviously)
Leave the source-code files, but remove the .sln, .csproj, and NAnt files so no-one could ‘accidentally’ compile it? (just a thought) From: Stephen Bohlen Sent: Thursday, August 18, 2011 1:06 PM To: [email protected] Subject: Re: [nhibernate-development] Re: GIT Conversion But I think the issue in that case would be that the SO reader (in this example) wouldn't ever get to SEE the readme file, they would just get an error when trying to resolve the link, right --? They would have to then be smart enough to explore the SVN repo in order to see the txt file and understand what's wrong (and how to fix it) -- probably not something most (any?) reader would really take the time to do. It seems that we're balancing two competing values here: 1. the desire to leave existing code in place so that anyone depending on it (links or otherwise) won't have broken references 2. the desire to ensure that anyone depending on this code *somehow* knows that what they are looking at is (potentially) obsolete Its not clear to me that its possible to accomplish both of these goals perfectly given what we have to work with. Is a compromise between these two competing values just to leave the existing code as-is and add the giant warning txt file? People linking to content will still get it resolved, anyone consuming/compiling the code would have to SEE the txt file for it to properly inform them of the problem. Other ideas ? Steve Bohlen [email protected] http://blog.unhandled-exceptions.com http://twitter.com/sbohlen On Thu, Aug 18, 2011 at 7:54 AM, Ramon Smits <[email protected]> wrote: On Thu, Aug 18, 2011 at 1:52 PM, Diego Mijelshon <[email protected]> wrote: There's a huge problem with that: it breaks links to files in the repo. I have more than once permalinked there when explaining stuff in stackoverflow, for example. Permalink to trunk? That sucks a little bit :-) still you could give a guide in the readme on how to resolve the link. -- Ramon
