Okay, quick (as in, writing this at midnight on Sunday straight into
django-admin without proofreading) weekly update before I fall asleep --
What I worked on this week
The good news: solved my worker-side woes that I mentioned in #5!
The bad news: it took me like ten hours of dead work and then I just ported everything to Heroku and got it working in thirty minutes.
I really, really wanted to like Elastic Beanstalk. It felt so promising, but it just could not deliver on the (relatively simple!) setup I needed. So I jumped over to Heroku, where things just worked a lot better. I'll eventually write about the setup, but all of the things that were a pain on Beanstalk had basically automagic solutions here --
phantomjs was just a buildpack,
django-rq was just a line in a Procfile, etc. etc.
I think Heroku's biggest mark against it was its pricing (and the implied admission of defeat, as if I'm saying I'm not smart enough for DevOps), but the truth is where I am right now I have a surplus of cash and a dearth of time. I spent around ten hours bashing my head against this stuff, and at least as a consultant I value my time at
$150/hr, so I spent $1500 trying and failing to get it set up on my own. Heroku, on the other hand, currently costs
$10/mo than my EBS set up. So it's like, rationally the math is fine for taking the L and just using Heroku -- its hard to fight the urge, though.
Even more than that, I think it's the mental toll of having run into a brick wall. It reminds me a lot of my first try at getting client/server sync working with Parse, and how a series of failures can really sap my enthusiasm for working on a code base. I want to avoid that happening again, so rather than doing the responsible thing and spending time getting
3.2.2 ready and out the door I'm gonna do a couple small tasks to build back up my cadence.
(Did some other minor stuff too: changed the
FullRecipeViewListController scrubber to coalesce number titles to "#"; added the ability to swipe rows in the
FavoriteRecipeViewListController to unfavorite; disabled interaction while syncing; and a slight redesign/unification of the web site.)
What I'm working on next week
Honestly, not that much. Freelance will probably be fairly busy so my evening hours are already fairly intense. Thankfully, I picked out seven easy/self-contained tasks that I can knock out, one a day:
- Close up shop at EBS and transfer anything relevant over
- Improve layout of
/ingredients-- not really crazy about the current design, it's just sorta meh.
- Fix issue where tapping the ingredients tab bar when you're in drill-down will bring the search bar wonky
- Remove error bashing (aka
try!) from the sync code, which should help with stability
- Migrate twitter API usage over to Buffer, since I have better things to do than writing out my own social media scheduler
- Tweet new recipes (this will hit buffer).
- Post errors to Slack
Nothing ambitious. Should be a good week, even if I feel guilty about pushing out the deadline for
v3.2.2 (which I do).
Haven't tried it yet, but discovered the Lazy Bear and oh my god it sounds delicious. Will probably have to wait til Friday though :(.