Yes, ideally our deployment process should be automated. Also, it should *not* be an SVN commit. It should be an rsync or an scp command. I would support any initiatives to move to either one of those. If we had automated deployment, this discussion would be moot.
How much would it cost us to just have a VPS with nginx? Switching to the topic of deployment docs now. Thanks, Shaz, for bringing this up in discussion. My opinion was that we shouldn't have impactful commands be copy-paste-able, which is why I had the instruction to commit in paragraph text. I think that if a committer doesn't read the full text of the deployment docs, *they should not be deploying*. I can see the argument that if they do read the text but just don't know *how* to commit in SVN, it's annoying to search. However at the top of that section is an explicit link to a quick SVN tutorial. I understand that not everyone reads the fine print, but IMO committers should, and we should explicitly discourage that behaviour. Ultimately I'm going to defer to Shaz here, but I think it's important to consider the benefits of making deployment *feel* more serious by making RTFD necessary. Kindly, Dmitry > On Sep 13, 2017, at 6:30 AM, Jan Piotrowski <piotrow...@gmail.com> wrote: > > I am actually surprised deploying is a manual process at all. > Having read the steps, I understand why of course. > > As a person that jumps in on all kinds of projects, I absolutely > prefer docs that list each and every little step needed (including all > the `cd` etc). > > The need for manual steps or checks could be emphasized by using a > numbered list or checklist for the individual steps. > > (Will this stay on SVN even after the repo switch to Github? Merging > and `gh-pages` is so nice and simple) > > -J > > 2017-09-13 9:02 GMT+02:00 Shazron <shaz...@gmail.com>: >> This relates solely to instructions on how to *build* the site, and not the >> contents of the site itself. >> >> Bringing this up here for discussion since a committer wants to revert a >> change by another committer, and there is potential for disagreement. >> >> The pull request to revert is here: >> https://github.com/apache/cordova-docs/pull/729 >> >> There has been discussion on the original change here: >> https://github.com/apache/cordova-docs/commit/96c5ab0f98c0b62160661dcd9a9db5549fe43f94 >> >> Two issues here: >> 1. The change from `gulp build --prod` to `npm run serve` >> 2. This instruction here (not reverted in the PR): >> https://github.com/apache/cordova-docs/commit/d61f3ddc84dac4b013c0607230b9cf10921a416b >> >> Issue (1) has some discussion in the GH link above for the original change. >> >> Issue (2) there was some discussion in the Cordova Slack, that the reason >> the `svn commit` wasn't there in the first place is to prevent copy/paste >> of the commands without going through the changes step by step since >> deploying to a site is an expensive operation that can screw up the site if >> proper care was not done. >> >> My reason for adding the command was that the instructions are not complete >> (when I had to do it myself when updating the docs for cordova-ios >> release). I understand the rationale, but the instructions seem incomplete >> (especially for new committers that haven't heard of SVN, I know they can >> Google it, but that's more friction) and my other reason is: we should >> trust that committers will do the right thing. >> >> Not to make a mountain out of a mole-hill but it's important that these >> revert discussions be out in the open so as misunderstandings/hurt feelings >> don't occur, and we can nip it in the bud. >> >> Thoughts from the community? > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@cordova.apache.org > For additional commands, e-mail: dev-h...@cordova.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@cordova.apache.org For additional commands, e-mail: dev-h...@cordova.apache.org
