Hey there,
I'm trying to figure out a strategy to get continuum building our
releases, got a few newbie questions:
1- Our software is mostly Java, but we have a few modules with C/C++
code in them. We have ant build scripts with custom ant tasks to build
those native pieces, no problem there. But we have multiple deployment
platforms (Linux/Solaris/HP-UX) and a few hardware/architectural
variations (Linux x86/ia64 etc.) where target artifacts of those native
modules should be built on the appropriate platform. So our build
process spans multiple hosts (with different OS/architectures), each
generating platform-specific artifacts and finally packaging them into a
single installable.
I'm not sure how to approach this problem in continuum terms --I suppose
we can install continuum on each build host (and perhaps share a common
database?), define a "master" build invoking ant, which in turn triggers
platform-specific builds via xml-rpc to each build server. Assuming
there's a actually a way to accomplish that, what would be the correct
approach?
2- We have a rather large number of products (30+), some of them have
huge codebases (few millions LOC --though I don't suppose that matters
to continuum) and almost all have multiple branches. Do those figures
pose a performance/scalability problem for continuum?
TIA,
Gokhan.
