On Aug 26, 2011, at 9:18 AM, Dennis E. Hamilton wrote: > Another place to put SVN dump files is on an [email protected] where > others can reach it. This is apparently one way to deliver to Apache > Infrastructure when the time comes. (Did you do that with the OpenOffice.org > Head, Rob?) > > Because of potentially separate projects having separate releases, it might > be useful to consider, > > /odf/simple/trunk/... > /odf/toolkit/trunk/... > /odf/subproject3/trunk/... > > and the peers of each trunk are releases or, if you prefer the /tag and > /branch structures that hold such things. (I favor releases being siblings > of trunk, but that's not typical.) > > My criterion is that /trunk/ and its peers are where the license and notice > files show up applicable to the particular releasable, > configurable/buildable/deployable component. > > Something I've been experimenting with that might be useful to you.
Is the plan to release separately for each component, or should we keep each component on its own release schedule? If it is separate then Dennis's structure makes more sense. If not then the odf/trunk/ makes more sense. Regards, Dave > > - Dennis > > > > > -----Original Message----- > From: Rob Weir [mailto:[email protected]] > Sent: Friday, August 26, 2011 06:49 > To: [email protected] > Subject: Re: List of code base needed to convert from Mercurial to SVN > > On Thu, Aug 25, 2011 at 11:14 PM, Devin Han <[email protected]> wrote: >> Hi all, >> >> I have converted the repository of Simple API from Mercurical to SVN with >> the help of the modified shell script I mentioned before. The other code >> bases will be converted later. > > Great! > > I can help review this if you upload the dumpfile. > > 1) svnadmin dump file:///local-repo > dumpfile > > 2) gzip or zip the dumpfile > > 3) Upload someplace, maybe to ODF Toolkit Union website > > 4) I can then download and load the dumpfile into a local respository. > > >> In order to avoid omit, it is necessary to make sure the code base list with >> your guys. >> > > Today we have a separate Hg repository per project, right? Or is it > possible to clone the entire project, via > https://hg.odftoolkit.org/hg/ ? > > In the end we want to end up with a single SVN repository at Apache. > > What will be easier to accomplish this: > > 1) Combine everything into a single Hg project and then convert to SVN? > > or > > 2) Convert separate Hg projects to SVN and then merge them together in SVN? > > Which makes it easier to preserve history? > > > In SVN our root will be: > > https://svn.apache.org/repos/asf/incubator/odf/ > > > So we'll want a directory structure for the trunk that looks like: > > /odf/trunk/component1 > /odf/trunk/component2 > /odf/trunk/component3, etc. > > [ ... ] >
