Jamie R Rytlewski
105 posts
Jamie R Rytlewski
@jamierytlewski
Senior App Dev
Okemos, MI Katılım Aralık 2006
324 Takip Edilen263 Takipçiler
2300+ builders at @UMich in Ann Arbor this week. The biggest @ClawCon yet.
Maize looks good on a lobster.
Some special things coming. Stay tuned.🦞
luma.com/clawconmichigan
English
@Dave_DotNet But saying indexed columns is kind of broad right? I could have an index on a column for states. Now would I want to use firstordefault or singleordefault probably not. But just saying indexed columns is not great advice.
English
#dotnet devs FYI ... there's no perceptible difference between FirstOrDefault and SingleOrDefault when used against indexed columns. If you're searching through non-indexed columns in large tables... well you've got a bigger problem, why aren't they indexed?
I like to use whatever gives the most accurate expression of intent to the dev reader based on what the result set should have. Using FoD on a primary key as a performance optimization, is gaining nothing but potentially confuses the reader ... why would I potentially get multiple results for a primary key query? 🤔
In memory collections are different especially in hot paths.
Mukesh Murugan@iammukeshm
Fact: Most .NET developers write EF Core queries that negatively impact database costs. Here are the TOP 5 EF Core mistakes killing your API performance right now: 1. No projections You return full entities when the client needs 3 fields. Fix: Use .Select(x => new { x.Id, x.Name, x.Status }) Result: Less data transferred, faster serialization, smaller SQL query. 2. N+1 queries You load orders, then loop and query each customer separately. Fix: Use .Include(o => o.Customer) or a projection that flattens the data. Result: 1 SQL query instead of 51. 3. Tracking read-only data EF Core tracks every entity by default. Change tracking costs memory and CPU. Fix: Add .AsNoTracking() to every read-only query. Result: 30-40% faster on large result sets. 4. Late filtering Calling .ToListAsync() before .Where() loads everything into memory first. Fix: Always filter, sort, and page BEFORE materializing. Result: SQL Server does the work instead of your app server. 5. SingleOrDefault vs FirstOrDefault SingleOrDefault scans the entire result set to guarantee uniqueness. Fix: Use FirstOrDefaultAsync when you just need the first match. Result: Faster queries, especially on large tables. None of these are hard to fix. All of them make a measurable difference. Check your codebase today. I guarantee at least 2 of these are hiding in there. Found this useful? Repost to your network. For more .NET tips, join my free newsletter: newsletter.codewithmukesh.com
English
@brankopetric00 Also v3 cache is EOL on February 1st
github.com/actions/cache
Eesti
GitHub Actions cache saved us $2,100/month.
The problem:
- 200+ workflow runs per day
- Each run: npm install (3 minutes)
- node_modules: 800MB
- Downloading dependencies every single time
The solution:
```yaml
- uses: actions/cache@v3
with:
path: node_modules
key: ${{ runner.os }}-node-${{ hashFiles('package-lock.json') }}
```
The impact:
- npm install: 3 minutes → 15 seconds
- Build time reduction: 40%
- Fewer billable minutes
We then added caching for:
- Docker layer cache (buildx cache)
- Terraform providers
- Go modules
- Python pip packages
Total monthly savings: $2,100
The cache is free (up to limits). The compute isn't.
If your CI downloads the same thing twice, you're burning money.
English
@brankopetric00 Are you not using the setup/node for npm install? It has caching built in and enabled.
github.com/actions/setup-…
English
@intentui @devongovett Again, make a PR. I know i made a PR for something and it was accepted. And we had a discussion on another about why it didn’t make sense and it was a good discussion. Hands down one of the better repos I’ve dealt with with PRs and issues.
English
@devongovett I’m not unwilling, but I truly believe you can make it far better. I’ve actually been waiting a very long time for this.
English
@CodifyBaseball does this include extra innings or is this only 9 inning games?
English
MLB games lasting 3+ hours:
2003: 28%
2004: 31%
2005: 30%
2006: 32%
2007: 39%
2008: 37%
2009: 38%
2010: 38%
2011: 40%
2012: 47%
2013: 52%
2014: 58%
2015: 46%
2016: 54%
2017: 62%
2018: 54%
2019: 62%
2020: 59%
2021: 67%
2022: 60%
2023: 17% 👀
2024: 12% 👀
2025: 15% 👀
English
@useTRMNL That’s crazy a firmware update did this. It was from someone not part of Trmnl right?
English
@devongovett I would expect the whole box to be clickable. Like a ghost button.
English
Which parts of this checkbox UI would you expect to be clickable? Only the box itself? Box and label? What about the description? Poll in next tweet. 🗳️
English
Jamie R Rytlewski retweetledi
If you know what a Danhausen is retweet this
English
English
@alexkehr Just curious as to how you know the strain and everything? Is that in the HealthKit?
English
Imagine telling a junior dev:
“To get the user’s name, just inject IHttpContextAccessor into your service.”
#dotnet has some wild hoops for the simplest things.
What’s the most unintuitive thing you’ve had to explain?
English
Jamie R Rytlewski retweetledi
@pankowecki Honestly the two biggest things is running and km. I run more than ride so the riding stats didn’t mean much to me. The km isn’t a big deal for me, but it might be for others. Otherwise it’s great. I can’t think of anything I’d like implemented right now. Ill have to see the stat
English
@jamierytlewski I won't be open sourcing yet but I am open to giving you access. What would you like to see implemented?
Would you like a setting to display in miles instead? Ideally TRMNL would send this setting but I can also add it myself.
English
@useTRMNL It’s so awesome that you have so many on back order and ready to ship. So excited and assembled in the USA.
English
@useTRMNL So not to be a pain and I love all what you are doing. But is there a way I can figure out how long I’m on the list for. I have so many ideas.
English
Many such cases
Evan Kapantais@evankapantais
TRMNL plugins are leaving me no time for anything else 😅
English
