TS-4048: Broken Vagrantfile in a lot of ways #490
Conversation
christoph-k
commented
Feb 20, 2016
- removed some broken boxes
- added some boxes
- changed from puppet-provisioning to shell-provisioning
Merge upstream changes
…o one with guest additions
|
This looks generally reasonable. I'd like the patch series to be a bit cleaner; maybe fix the vagrant box URLs in the first commit, then switch over to the shell provisioner. We need to retain the additional build packages ( |
|
Yeah i know what you mean about that patch series... working on it I just checked for configure to run in default configuration when looking for those dependencies trying to keep the images small. I can add those optional dependencies as well. Would be great if they were documented as optional though (doclink). What do you think about dependencies for plugins? Shall they be added as well? That FreeBSD image is gone and there wasn't any provisioning at all so that probably never worked ootb anyways. If that's still needed i'll try to add it as well but i'd prefer to do that after all the other stuff is working again. |
thanks :)
The original goal of these was so that dev could spin up build environments for configurations they didn't have easy access to, which meant that I installed all the dependencies to build all the optional components. If you can get all the dependencies in, that would be great. On RHEL-based systems I remember it was a bit tedious to get the right dependencies out of EPEL though.
No that's fine; I just wanted to check whether this was a deliberate decision. A FreeBSD image will just bitrot again unless there's someone to take care of it. |
|
@jpeach will shepard. |
|
I just added those optional dependencies. Cleaning up that patch series is pretty hard though. Looks more like a complete rewrite than a couple of fixes. I'm still thinking about how to split all those changes without leaving it in an even worse state. Those commits work alltogether... applying only some of them does'nt really make any sense. Maybe just squash em? |
|
Thanks @christoph-k86! |
* Revert "Fix issue with plugin supplied address - suppress DNS lookup in that case. (apache#486)" This reverts commit 5023a7d. * Revert "Fix: avoid reusing null hostent if the DNS query failed. (apache#485)" This reverts commit 85be058. * Revert "Revert API change - maintain 9.x compatibility. (apache#478)" This reverts commit 26a7fc9. * Revert "Add method to write an IpAddr value to a sockaddr. (apache#7821)" This reverts commit 4a5ec19. * Revert "DNS: Clean up argument passing to DNS queries. (apache#7778)" This reverts commit bc546e4. * Revert "Add overload for memcpy to take a destination buffer and source string_view / TextView (apache#7732)" This reverts commit e782f7b. * Revert "Add basic type aliases for std::chrono types to ink_time.h for future use. (apache#7482)" This reverts commit 9554c05. * Revert "Add Au test for strategies.yaml, with consistent hashing, with fallover. (apache#7914) (apache#480)" This reverts commit 123b53f. * Revert "HostDB restructuring." This reverts commit dae938f.
* Resource Constraints: schedule reloading configs * Remove mutex * Start RM with accepts, make reconfigure synchronous at startup * Update test/stub files * Update unit test env * Removed wrapper script