Subject: Re: [boost] [GSoC] Bonding period
From: Stefan (mstefanro_at_[hidden])
Date: 2010-05-03 14:25:42
Andrew Sutton wrote:
>> I have already created my directory structure in
>>> /sandbox/SOC/2010/stringalgos/ but I assume it would have been better to
>>> branch from the beginning.
>> Is this double-branching the right solution here? If not, let me know how I
>> should proceed.
> It might be easier just to start in sandbox and then merge or patch that as
> you go. In the end, it's really up to you determine what works the best for
> you -- unless your mentor has different ideas :)
> Andrew Sutton
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Thanks for the reply, I will discuss this with my mentor soon.
However, although I'm not that experienced with SVN, I don't see why not
use branching here.
At some point in time, I will need to modify some files which already
exist in trunk's version of the library (headers, documentation,
example, tests), which means I will have to SVN COPY those files from
trunk to my sandbox directory. Since I will do that for each file
individually, I will also have to periodically do SVN MERGE for each
copied file (in case the trunk version of the file gets modified while
I'm also modifying it in my sandbox), which sounds like some extra work.
Unless I branch the old lib, I will also have to keep track of which
files have I SVN COPIED and which are mine (so I will know which ones
need to be merged). If I had it as a branch, one single merge from trunk
would make all my files in sandbox up-to-date with respect to trunk.
It also allows me to have all the library's files in the same directory
(all documentation, all tests, examples etc.) which makes it easier for
me to work with them and compile documentation/tests/examples.
Additionally, hypothetically, I wouldn't be able to delete a file from
the old library if it's not needed in this newer version unless I
actually had a full branch of the old library.
As mentioned above, I will discuss this with my mentor more thoroughly
(I believe my progress will be just as easily monitorable even if
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk