Hacker Timesnew | past | comments | ask | show | jobs | submit | chdsbd's commentslogin

With regards to commits, I think it’s better to _not_ worry about making them nice. I think time is better spent making the PR description clear.

Make as many commits as you want, just “Squash and merge” at the end using your pull request body for the commit message.


Please no. Squash and merge means that when someone (maybe you) comes across a strange-looking line of code in 18 months and uses git blame to figure out why it was added, they'll be presented with a 1500-line wall of text with no context.

Commits are a tool. Learn to use them.


Hello HN,

Kodiak is an open source GitHub app that can update, approve, and merge pull requests.

At work we were getting frustrated waiting for CI to merge pull requests, so I built Kodiak to automate this tedious work.

With Kodiak you don't need to wait for CI checks to finish, you simply add a label and Kodiak will merge your pull request when ready. It's super helpful when you have long running CI or "Require branches to be up to date before merging" configured via GitHub branch protection.

It's really simple to get started: https://kodiakhq.com/docs/quickstart

github repo: https://github.com/chdsbd/kodiak


A bit of shameless self promotion: I built a more basic merge bot for GitHub that efficiently updates and merges PRs because we were wasting a ton of time keeping branches updated at work.

https://github.com/chdsbd/kodiak


We started using Kodiak after the Auto Rebase bot was discontinued a few weeks ago. Other than confusing me with a normal merge when I thought it had been configured to do squash-merges, it works great. Thanks for releasing it!


Thanks for the chart comparing it to many alternatives. It both makes it clear what new things Kodiak does and what alternatives one could look at.


It seems that MUMPS doesn't have serializable isolation.


Yeah, when I read about that I thought it sounded neat—make sure the index updates with the data. On the other hand, I think of CAP theorem as an iron triangle. If you are gaining consistency, what’s the trade off?


This _is_ the usual trade off, but what makes FoundationDB so crazy is that it's a CP system that has a performance profile that AP systems would have a hard time matching.


TablePlus supports Postgres/MySQL/Sql Server and others. It has a paid version which adds a few more features, but the free version works fine for me.

https://tableplus.io/


I can't seem to collapse comments on the site.


+1 for ios.

In addition, it feels slower than regular HN. No actual benchmarks though.

Checking https://github.com/clintonwoo/hackernews-react-graphql/blob/... it might be on purpose though. Toggle code is commented.


How did you handle upgrading between major Postgres versions with Docker when the container only has one version (you need both for major version updates)?

There is an issue about this (from 2014), but it has yet to be resolved: https://github.com/docker-library/postgres/issues/37


I can't remember if we ever did this, but the solution seems straightforward, just start an upgrade container that has both versions.

We definitely did not use the container that's in the locker library though. Ours did some custom checks on startup.


You still need an Apple developer account to sign and run an open source extension


I was taught in ~5th grade that you can trust a ".org, .gov, or .edu" website when looking for sources. (This was in a Massachusetts public school)


Did .org ever have any oversight?


not exactly - but your used to be able to more-or-less count on it being owned/operated by a non-profit of some kind


I'm curious. Why do you say this would make cheating easier?


For a cheater, these are lucrative areas: Areas with low income and education rates, where the populace doesn't typically vote. Stuffing the box won't go noticed due to low election turnout.

In Kansas, we have a system to prevent ballot stuffing. You must show state or federal issued ID. Your name must be in the pollbook for the location and pre-registered before election day.

They have two areas: The first, you show your ID and they check your entry in the location's pollbook. The fact that you showed up to vote along with the ID you presented is recorded electronically statewide and is public information instantly.

They then print an anonymous "voting ticket" and you're taking to a separate area or room. A smart card is exchanged for your voting ticket is put into an air-gapped system and you cast your ballot. Your smart card is removed and tallied on a third air-gapped system. When the poll closes, the tally machine is handed over to the county.

This system makes it nearly impossible to stuff the box, because the number of votes cast at a location cannot exceed the number of IDs presented at the checkin location. It doesn't protect from changing votes, which would require a much more sophisticated attack on the machines themselves.

It's not perfect, but it is fairly good at stopping unsophisticated opportunistic attackers. I would feel much better if all of the USA adopted such a protocol if we go to a popular vote, but that's unlikely to happen, so I can't say I'm comfortable moving away from the electoral college either.


Wait, what states don't use election rolls (id checks do vary by state)?

I'm a lot happier with paper ballots + electronic tabulation than I am with electronic ballots, what do you think those parts of the process you describe are adding?


Right now it's nearly impossible to clean election rolls of dead etc. voters, e.g. the Feds fight this hard.

However, state and Federal issued IDs have a subsidiary feature, perhaps even more important that on the spot election verification now that I think about it, in that they expire, and I'm sure some of them also have revocation measures after the ID holder dies. This is much more true for people moving.

Agreed on the paper ballot + electronic counting, and very glad my county in neighboring Missouri uses that system, I never trusted the totally electronic or electromechanical systems I used in Arlington, VA or Brookline, MA


Many many fewer precincts and political domains.

The US has a total of 3,143 counties or equivalents per Wikipedia, obviously divided between 50 states and D.C., each a separate political domain, both counties and states. Corrupting enough of those under the Electoral College's constraints is a lot harder.


Because you're going directly off of the vote, instead of a representation of that vote via the elector for that area. It's a subtle difference, but one that brings to light some of the fears with direct election.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: