- 25 Jul, 2020 3 commits
-
-
Georges Racinet authored
This should help bring the confusion down for this simple class of mistakes.
-
Georges Racinet authored
The heptapod-0-14 branch of the HDK is about providing a setup for the current stable Heptapod series (branch `heptapod-stable` of the Rails application) and *does not work* for the default series because of the bigger setup changes involved by the bump from GitLab 12 to 13.
-
Georges Racinet authored
-
- 19 Jul, 2020 2 commits
-
-
Georges Racinet authored
With our functional test suite performing various SSH tests, it's a pain for developers creating new workspaces regularly not to have it enabled by default.
-
Georges Racinet authored
With Heptapod 0.15, we move to GitLab 13, which incurs a significantly different setup. It does not seem worth the effort to change the current default HDK branch to adapt to Heptapod 0.14, so let's have a branch. It's likely we'll have something similar at the next major version change, but we have a long time before that.
-
- 13 Jul, 2020 1 commit
-
-
Georges Racinet authored
-
- 11 Jul, 2020 1 commit
-
-
Georges Racinet authored
-
- 09 Jul, 2020 1 commit
-
-
Georges Racinet authored
With the Rails app at 99d9d4c772ce, the seeding of testing projects work, which lets the whole build finish normally
-
- 08 Jul, 2020 4 commits
-
-
Georges Racinet authored
-
Georges Racinet authored
This is intended to replace `gdk init`, but starting from within a HDK checkout. Needed because the first thing that `gdk init` does is a Git clone of the GDK main repo. We had to depend on `.ruby-version` so that gdk itself can run. This doesn't solve all chicken-and-eggs problem: we still depend on `rake` being available to produce `gdk-config.mk`.
-
Georges Racinet authored
Currently, postgresql-seed-rails fails on the the 'projects' step. That wouldn't prevent developping Heptapod in itself, as long as the `heptapod` target is completed first. This artificial dependency can be removed once the seeding step succeeds, even if we simply make it smaller.
-
Georges Racinet authored
From the new `heptapod_series` setting in `gdk.yml` (defaulting to `default), a consistent list of revisions is derived in `config.rb`, dumped in the `gdk-config.mk` for final usage in `Makefile`. For now, since we don't have update capabilities, this will have to be set prior to the first call to `make`. The `stable` series is 0.13.x, a state of affairs that shouldn't stay for very long.
-
- 07 Jul, 2020 8 commits
-
-
Georges Racinet authored
derived from `.gitignore` with conversion of `/` patterns into rootglobs and addition of our own generated content. Probably a bit crude, does the job for now
-
Georges Racinet authored
-
Georges Racinet authored
This completes making an entirely functional Heptapod Development Kit, although some things are still hardcoded.
-
Georges Racinet authored
-
Georges Racinet authored
This references the `hg` executable and the Mercurial configuration in the Rails app and Heptapod Shell configuration files. The `gunicorn-heptapod` script, the HTTP serving component for Mercurial content, wraps the `gunicorn` executable with the proper `HGRC_PATH`
-
Georges Racinet authored
Up to now, Go 1.13 has been working well enough for me for Heptapod development. Of course, switching to 1.14 will be advised, since the GDK insists on it, but that doesn't seem to be really necessary right now, we'll do it once the rest is ready.
-
Georges Racinet authored
The new `py-heptapod` make target takes care of all of that. Then, of course, we'll need to generate the appropriate configurations to use it.
-
Georges Racinet authored
We start from a fairly recent GDK version and change the source URLs for the Rails app, Heptapod Shell and Heptapod Workhorse. For now, we won't rename the cloned components to the more logical and clearer `heptapod`, `heptapod-shell` and `heptapod-workhorse`, because `gitlab/` is hardcoded a lot in the Makefile. And frankly, there are enough reasons for failure as it is: we'll do these renames only after we have a fully working kit. All Mercurial cloning is done through a share pool, whose path is for now hardcoded to the `hg-share-pool` sibling of the GDK root. We removed all `update` capabilities (will maybe make them work later) and have the `heptapod` (hardcoded) branch of each component checked out. Apart from that, we keep roughly the same logic as upstream GDK. At this point, `make` works up to the creation of the `root` user (`seed_fu`) and fails as expected on the creation of test projects - something we'll look at later. The Rails app starts, but the `Rails.root` had to be changed to just `root` in `lib/gitlab.rb` for background jobs to start, which means we may have something deeper to correct here. Of course, without the Python components (the next step), we won't go too far.
-
- 01 Jun, 2020 5 commits
-
-
Ash McKenzie authored
Support full SHAs for components See merge request gitlab-org/gitlab-development-kit!1297
-
Ash McKenzie authored
-
Tiger Watson authored
Ensure gitaly has its Ruby gems installed See merge request gitlab-org/gitlab-development-kit!1296
-
Ash McKenzie authored
-
Ash McKenzie authored
Add option to disable sourcemaps for webpack See merge request gitlab-org/gitlab-development-kit!1289
-
- 29 May, 2020 1 commit
-
-
Mike Greiling authored
-
- 01 Jun, 2020 1 commit
-
-
Ash McKenzie authored
Require Git v2.24.0 for the GDK See merge request gitlab-org/gitlab-development-kit!1287
-
- 29 May, 2020 1 commit
-
-
Stan Hu authored
Gitaly added the use of `--end-of-options` in https://gitlab.com/gitlab-org/gitaly/-/merge_requests/2192, but this was introduced in Git v2.24.0: https://github.com/git/git/blob/2d5e9f31ac46017895ce6a183467037d29ceb9d3/Documentation/RelNotes/2.24.0.txt#L18-L23
-
- 01 Jun, 2020 4 commits
-
-
Ash McKenzie authored
Fix TEST_IMAGE -> VERIFY_IMAGE rename miss See merge request gitlab-org/gitlab-development-kit!1295
-
Ash McKenzie authored
-
Ash McKenzie authored
update ruby version to 2.6.6 See merge request gitlab-org/gitlab-development-kit!1290
-
David Flores authored
-
- 29 May, 2020 8 commits
-
-
Stan Hu authored
Make Praefect DB seeding a prerequisite for Rails DB seeding Closes #942 See merge request gitlab-org/gitlab-development-kit!1288
-
Stan Hu authored
Now that Praefect requires the database to be set up, we should seed the Praefect database first so that `support/bootstrap-rails` can run properly. Closes https://gitlab.com/gitlab-org/gitlab-development-kit/-/issues/942
-
Toon Claes authored
Rename config.local_hostname to config.listen_address Closes #938 See merge request gitlab-org/gitlab-development-kit!1283
-
Ash McKenzie authored
-
Toon Claes authored
Sort .gitignore See merge request gitlab-org/gitlab-development-kit!1286
-
Ash McKenzie authored
-
Ash McKenzie authored
-
Ash McKenzie authored
Restructure GDK documentation See merge request gitlab-org/gitlab-development-kit!1280
-