>> On Thu, October 28, 2010 00:41, Dan Haywood wrote: >> > It doesn't look like you've done a full svn update? >> >> That's a question - My first svn update got interrupted (link forcibly >> closed by host). The second update fetched more files. The third says >> update complete. >> >> How error prone is svn when links are forcibly closed? I thought it was >> supposed to be atomic. >> > > Commits are atomic, but doing an svn update isn't, as you've seen.
See below.. >> >> If I do have an unstable svn state, what can repair it? > > Not sure what you mean by "unstable svn state"; an interrupted svn update > is not serious, just try again. > This is just the problem - you're suggesting that I don't have a fully up-to-date update. But I repeatedly performed updates until the dialog stated that I was up-to-date. So, I believe, it would be reasonable to expect that I am fully synchronized. Therefore, any issues that I am now experiencing are either caused by my svn state being unstable (i.e. not a reliable mirror of the server) or uncommitted differences from source (i.e. you have not comitted all changes). A third option is that somehow your build environment is solving dependencies that mine isn't. I guess one of my questions is: Can I somehow compute (for example) an md5 checksum of all sources and compare it against a server value? If done on a per-module basis, then I can at least determine which modules of mine are not properly synchronised with the server? FYI: I am building from command line.
