Also thinking about this from the POV  of the fact that the existing SVN
repo URL is widely-known/posted/linked-to/etc.  With this kind of an
approach, anyone blindly doing an svn update would get the new file and
(hopefully!) notice it in their working copy...

Steve Bohlen
[email protected]
http://blog.unhandled-exceptions.com
http://twitter.com/sbohlen


On Wed, Aug 17, 2011 at 5:11 PM, Stephen Bohlen <[email protected]> wrote:

> Since the existing SVN repo is effectively read-only for everyone but the
> small group of committers, my guess is that we're "safe" just ensuring that
> the existing committers are told "hands off the SVN repo for the foreseeable
> future" but I'm also thinking of how someone new to NH would know that the
> SVN repo was obsolete (and this not to use it as the basis for creating
> patches, etc.).
>
> What about committing a text file into the root of the SVN repo named
> THIS_REPOSITORY_IS_OBSOLETE___DO_NOT_USE___READ_ME_FOR_DETAILS.txt that
> explains the location of the new github repo so that there's at least some
> way of new folks not making the mistake of working with this repo any
> longer?
>
>
> Steve Bohlen
> [email protected]
> http://blog.unhandled-exceptions.com
> http://twitter.com/sbohlen
>
>
> On Wed, Aug 17, 2011 at 5:02 PM, Patrick Earl <[email protected]> wrote:
>
>> On Wed, Aug 17, 2011 at 1:26 PM, Stephen Bohlen <[email protected]>
>> wrote:
>> > I think that there are probably several housekeeping issues that still
>> > remain to be addressed around this process:
>> >
>> > We need to decide what to do about the disposition of the existing SVN
>> repo
>> > (shut it down completely, annotate its content as SUPERSEDED and direct
>> ppl
>> > to the git repo via a readme note in it, etc., etc.).  What are people's
>> > ideas about how best to handle this?  Is it the case at this point that
>> the
>> > entire contents of the SVN repo are in git (e.g., there have been no
>> > subsequent commits to the SVN repo since the import to git)?
>>
>> There have been no new commits and all branches and tags have been
>> imported to the best of my knowledge.  6004 was the last SVN commit
>> included.  There have been a couple commits on the GIT repository
>> after the import.
>>
>> I believe we should start directing people to the new repository, and
>> if possible change the existing one to read-only.  I don't think we
>> should delete it until more time has passed.
>>
>> We should decide on our suggested workflow for issues / pull requests
>> / feature branches.  The castle project folks might be able to provide
>> some insight here, since they went though a similar situation.
>>
>> > We need to decide how to handle the existing SVN-based patches that have
>> > been attached to still-open JIRA issues.  Properly applying these to
>> TRUNK
>> > as the contents of the SVN and git repos diverge over time is probably
>> going
>> > to be a non-trivial process (unless someone has some pearls of wisdom to
>> > share re: tricks for handling this simply...I've never actually tried to
>> > apply an SVN patch to a non-SVN-controlled set of files...is this even
>> > possible???)
>>
>> The SVN patches are almost just unified diffs.  These can be applied
>> without much effort using generic patch tools.
>>
>> For example:
>> http://stackoverflow.com/questions/659467/how-to-apply-svn-diff-to-git
>>
>> The patches in JIRA tend to be small or require significant
>> modification anyways, so I don't think this will be a major issue in
>> practice.
>>
>>         Patrick Earl
>>
>
>

Reply via email to