on Wed Dec 26 2012, Rene Rivera <grafikrobot-AT-gmail.com> wrote:
On 12/25/2012 3:08 PM, Dave Abrahams wrote:
on Tue Dec 25 2012, Rene Rivera <grafikrobot-AT-gmail.com> wrote:
After looking at both of those I see no point in using the github api (or additional structure data from sublime -- not totally sure where the submodule info comes from in this case though) for this as it provides no additional information than one can get from just parsing the ".gitmodules" file.
I'm pretty sure that's not correct. The .gitmodules file doesn't contain information about which commit to check out for each submodule.
Right it doesn't. But your ryppl code doesn't handle that either since it fetches the repos individually from the non-version-specific master branches (AFAICT).
No, that's also incorrect. Ryppl uses zeroinstall to fetch specific versions of dependencies based on the output of a statisfiability solver.
Hence the complexity of supporting testing with ZIPs is now a magnitude larger as it means dealing with fetching more than a hundred individual repos :-(
Which now seems the only choice. At the tester side I will have to get the boost-master archive. Then parse out the ".gitmodules" file. And get each subrepo archive individually. Which increases the likelihood of failure considerably.
I'm not sure. Isn't it true that shorter transfers are more likely to succeed than longer ones?
Perhaps, if one happens to have an not reliable internet connection. But I would expect testers to have reliable connections. But that's a minor unreliability.. The more likely problem is in code bugs in the testing script ;-)
OK, any new code comes with a risk of bugs, but seriously: this is not the kind of monstrous increase in complexity that will be difficult to code for, or to debug. -- Dave Abrahams BoostPro Computing Software Development Training http://www.boostpro.com Clang/LLVM/EDG Compilers C++ Boost