316 Matching Annotations
  1. Last 7 days
    1. Collecting per-second measurements of CPU load might yield interesting data, but such frequent measurements may be very expensive to collect, store, and analyze.

      Revisit the log files on our production server.

    1. it is better to allow an error budget—a rate at which the SLOs can be missed—and track that on a daily or weekly basis
    2. To save effort, build a set of reusable SLI templates for each common metric; these also make it simpler for everyone to understand what a specific SLI means.
    1. With over 16 million pulls per month, Google’s `distroless` base images are widely used and depended on by large projects like Kubernetes and Istio. These minimal images don’t include common tools like shells or package managers, making their attack surface (and download size!) smaller than traditional base images such as `ubuntu` or `alpine`.

      I need to check these out.

  2. Apr 2021
    1. There's nothing to stop you from doing initializer code in a file that lives in app/models. for example class MyClass def self.run_me_when_the_class_is_loaded end end MyClass.run_me_when_the_class_is_loaded MyClass.run_me... will run when the class is loaded .... which is what we want, right? Not sure if its the Rails way.... but its extremely straightforward, and does not depend on the shifting winds of Rails.

      does not depend on the shifting winds of Rails.

    1. Of course you must not use plain-text passwords and place them directly into scripts. You even must not use telnet protocol at all. And avoid ftp, too. I needn’t say why you should use ssh, instead, need I? And you also must not plug your fingers into 220 voltage AC-output. Telnet was chosen for examples as less harmless alternative, because it’s getting rare in real life, but it can show all basic functions of expect-like tools, even abilities to send passwords. BUT, you can use “Expect and Co” to do other things, I just show the direction.
    2. But in all this incongruous abundance you'll certanly find the links to expect It's just what is wanted: the tool, which is traditionally used to communicate automatically with interactive programs. And as it always occurs, there is unfortunately a little fault in it: expect needs the programming language TCL to be present. Nevertheless if it doesn't discourage you to install and learn one more, though very powerful language, then you can stop your search, because expect and TCL with or without TK have everything and even more for you to write scripts.
    1. “Who cares? Let’s just go with the style-guide” — to which my response is that caring about the details is in the heart of much of our doings. Yes, this is not a major issue; def self.method is not even a code smell. Actually, that whole debate is on the verge of being incidental. Yet the learning process and the gained knowledge involved in understanding each choice is alone worth the discussion. Furthermore, I believe that the class << self notation echoes a better, more stable understanding of Ruby and Object Orientation in Ruby. Lastly, remember that style-guides may change or be altered (carefully, though!).
    1. We strongly encourage you to reset any personal access tokens and OAuth tokens you have.

      Reset my access tokens. Pretty sure I have at least one.

    1. This approach is preferable to overriding authenticate_user! in your controller because it won't clobber a lot of "behind the scenes" stuff Devise does (such as storing the attempted URL so the user can be redirected after successful sign in).
    1. # +devise_for+ is meant to play nicely with other routes methods. For example, # by calling +devise_for+ inside a namespace, it automatically nests your devise # controllers: # # namespace :publisher do # devise_for :account # end
  3. Mar 2021
    1. This should link to / explain the relationship to: https://en.wikipedia.org/wiki/Class_(computer_programming) (which I believe is a way of expressing / codifying semantic classes into source code).

      It should also link to / explain the relationship to: https://en.wikipedia.org/wiki/Type_theory

    2. (Not answered on this stub article)

      What, precisely, is the distinction/difference between a semantic class and a semantic field? At the very least, you would say that they are themselves both very much within the same semantic field.

      So, is a semantic class distinct from a semantic field in that semantic class is a more well-defined/clear-cut semantic field? And a semantic field is a more fluid, nebulous, not well-defined field (in the same sense as a magnetic field, which has no distinct boundary whatsoever, only a decay as you move further away from its source) ("semantic fields are constantly flowing into each other")?

      If so, could you even say that a semantic class is a kind of (hyponym) of semantic field?

      Maybe I should pose this question on a semantics forum.

    1. a Docker container running a very simple NodeJS web server with the Graphile library (and some additional Netflix internal components for security, logging, metrics, and monitoring) could provide a “better REST than REST” or “REST++” platform for rapid development efforts

      Give this a try.

      1. Get out the city shapefile and overlay these values.
      2. Get the most current version of the voter registration database.
      3. Determine the number of voters who, according to the Spokane Journal of Business, took part in the 2018 school bond vote, but not the stadium advisory vote. (Geographically, these would be folks who live within SD81, but outside of the city limits.)
    1. Meh... as I said earlier, I think using Webpack is the recommended way now. Another issue is there is no way to generate source maps in production.
    2. But yeah, I'm not sure how you would determine which was the "recommended way" really. I don't see anything in Rails docs saying either way.
    3. But last I have seen comments from DHH, he considered webpack(er) recommended for JS, but Sprockets still the preferred solution for (S)CSS.
    4. Is there a PR to... something? sassc-rails? That would make the patch not necessary? (I don't know if there's any good way to monkey-patch that in, I think you have to fork? So some change seems required...) Should the defaults be different somehow? This is very difficult to figure out.
    1. Beykat yi duñu dem tool altine.

      Les cultivateurs ne vont pas au champ le lundi.

      beykat bi -- farmer 👩🏾‍🌾 (from bey -- to farm/cultivate).

      yi -- the (indicates plurality).

      duñu -- do not/no one (?).

      dem v. -- to go, leave, etc.

      tool bi -- field, orchard.

      altine ji -- (Arabic) Monday.

  4. Feb 2021
    1. URI::MailTo::EMAIL_REGEXP

      First time I've seen someone create a validator by simply matching against URI::MailTo::EMAIL_REGEXP from std lib. More often you see people copying and pasting some really long regex that they don't understand and is probably not loose enough. It's much better, though, to simply reuse a standard one from a library — by reference, rather than copying and pasting!!

    1. For branching out a separate path in an activity, use the Path() macro. It’s a convenient, simple way to declare alternative routes

      Seems like this would be a very common need: once you switch to a custom failure track, you want it to stay on that track until the end!!!

      The problem is that in a Railway, everything automatically has 2 outputs. But we really only need one (which is exactly what Path gives us). And you end up fighting the defaults when there are the automatic 2 outputs, because you have to remember to explicitly/verbosely redirect all of those outputs or they may end up going somewhere you don't want them to go.

      The default behavior of everything going to the next defined step is not helpful for doing that, and in fact is quite frustrating because you don't want unrelated steps to accidentally end up on one of the tasks in your custom failure track.

      And you can't use fail for custom-track steps becase that breaks magnetic_to for some reason.

      I was finding myself very in need of something like this, and was about to write my own DSL, but then I discovered this. I still think it needs a better DSL than this, but at least they provided a way to do this. Much needed.

      For this example, I might write something like this:

      step :decide_type, Output(Activity::Left, :credit_card) => Track(:with_credit_card)
      # Create the track, which would automatically create an implicit End with the same id.
      Track(:with_credit_card) do
          step :authorize
          step :charge

      I guess that's not much different than theirs. Main improvement is it avoids ugly need to specify end_id/end_task.

      But that wouldn't actually be enough either in this example, because you would actually want to have a failure track there and a path doesn't have one ... so it sounds like Subprocess and a new self-contained ProcessCreditCard Railway would be the best solution for this particular example... Subprocess is the ultimate in flexibility and gives us all the flexibility we need)

      But what if you had a path that you needed to direct to from 2 different tasks' outputs?

      Example: I came up with this, but it takes a lot of effort to keep my custom path/track hidden/"isolated" and prevent other tasks from automatically/implicitly going into those steps:

      class Example::ValidationErrorTrack < Trailblazer::Activity::Railway
        step :validate_model, Output(:failure) => Track(:validation_error)
        step :save,           Output(:failure) => Track(:validation_error)
        # Can't use fail here or the magnetic_to won't work and  Track(:validation_error) won't work
        step :log_validation_error, magnetic_to: :validation_error,
          Output(:success) => End(:validation_error), 
          Output(:failure) => End(:validation_error) 
      puts Trailblazer::Developer.render o
       {Trailblazer::Activity::Right} => #<Trailblazer::Activity::TaskBuilder::Task user_proc=validate_model>
      #<Trailblazer::Activity::TaskBuilder::Task user_proc=validate_model>
       {Trailblazer::Activity::Left} => #<Trailblazer::Activity::TaskBuilder::Task user_proc=log_validation_error>
       {Trailblazer::Activity::Right} => #<Trailblazer::Activity::TaskBuilder::Task user_proc=save>
      #<Trailblazer::Activity::TaskBuilder::Task user_proc=save>
       {Trailblazer::Activity::Left} => #<Trailblazer::Activity::TaskBuilder::Task user_proc=log_validation_error>
       {Trailblazer::Activity::Right} => #<End/:success>
      #<Trailblazer::Activity::TaskBuilder::Task user_proc=log_validation_error>
       {Trailblazer::Activity::Left} => #<End/:validation_error>
       {Trailblazer::Activity::Right} => #<End/:validation_error>

      Now attempt to do it with Path... Does the Path() have an ID we can reference? Or maybe we just keep a reference to the object and use it directly in 2 different places?

      class Example::ValidationErrorTrack::VPathHelper1 < Trailblazer::Activity::Railway
         validation_error_path = Path(end_id: "End.validation_error", end_task: End(:validation_error)) do
          step :log_validation_error
        step :validate_model, Output(:failure) => validation_error_path
        step :save,           Output(:failure) => validation_error_path
      o=Example::ValidationErrorTrack::VPathHelper1; puts Trailblazer::Developer.render o
       {Trailblazer::Activity::Right} => #<Trailblazer::Activity::TaskBuilder::Task user_proc=validate_model>
      #<Trailblazer::Activity::TaskBuilder::Task user_proc=validate_model>
       {Trailblazer::Activity::Left} => #<Trailblazer::Activity::TaskBuilder::Task user_proc=log_validation_error>
       {Trailblazer::Activity::Right} => #<Trailblazer::Activity::TaskBuilder::Task user_proc=save>
      #<Trailblazer::Activity::TaskBuilder::Task user_proc=log_validation_error>
       {Trailblazer::Activity::Right} => #<End/:validation_error>
      #<Trailblazer::Activity::TaskBuilder::Task user_proc=save>
       {Trailblazer::Activity::Left} => #<Trailblazer::Activity::TaskBuilder::Task user_proc=log_validation_error>
       {Trailblazer::Activity::Right} => #<End/:success>

      It's just too bad that:

      • there's not a Railway helper in case you want multiple outputs, though we could probably create one pretty easily using Path as our template
      • we can't "inline" a separate Railway acitivity (Subprocess "nests" it rather than "inlines")
    2. step :direct_debit

      I don't think we would/should really want to make this the "success" (Right) path and :credit_card be the "failure" (Left) track.

      Maybe it's okay to repurpose Left and Right for something other than failure/success ... but only if we can actually change the default semantic of those signals/outputs. Is that possible? Maybe there's a way to override or delete the default outputs?

    3. Patching has no implicit, magical side-effects and is strongly encouraged to customize flows for a specific case in a quick and consise way.
    4. While you could nest an activity into another manually, the Subprocess macro will come in handy.
    5. The macro automatically wires all of Validate’s ends to the known counter-part tracks.
    1. They do not maintain a to-do list (mentally or physically).
    2. If you ask my former students, they will tell you that as a teacher, my goal is to do nothing. I dream of the day when I can sit at my desk, feet propped up, reading a book, while the classroom bursts with activity and learning around me.
    1. While Trailblazer offers you abstraction layers for all aspects of Ruby On Rails, it does not missionize you. Wherever you want, you may fall back to the "Rails Way" with fat models, monolithic controllers, global helpers, etc. This is not a bad thing, but allows you to step-wise introduce Trailblazer's encapsulation in your app without having to rewrite it.
    1. ActiveModel provides a powerful framework for defining callbacks. ActiveInteraction hooks into that framework to allow hooking into various parts of an interaction's lifecycle.
    1. Personal todo lists don’t depend on others using the same system (no network effects)

      They don't unless you're building a wiki or commonplace book that can interact with those of others. (Roam research isn't doing this---yet, but they should.) Ideally small building block pieces will allow it to dovetail with other systems that could potentially do the same thing.

    1. {a: 1, b: 2, c: 3, d: 4} => {a:, b:, **rest} # a == 1, b == 2, rest == {:c=>3, :d=>4}

      equivalent in javascript:

      {a, b, ...rest} = {a: 1, b: 2, c: 3, d: 4}

      Not a bad replacement for that! I still find javascript's syntax a little more easily readable and natural, but given that we can't use the same syntax (probably because it would be incompatible with existing syntax rules that we can't break for compatibility reasons, unfortunately), this is a pretty good compromise/solution that they've come up with.

    1. Examples of different ways of defining forms

      Wow, that's a lot of different ways.

      The inline_form way in particular seems interesting to me, though it's worth noting that that method is just an example, not actually part of this project's code, so it's not really a first-class option like the other options.

    1. DSLs can be problematic for the user since the user has to manage state (e.g. am I supposed to call valid? first or update_attributes?). This is exactly why the #validate is the only method to change state in Reform.
    2. The reason Reform does updating attributes and validation in the same step is because I wanna reduce public methods. This is to save users from having to remember state.

      I see what he means, but what would you call this (tag)? "have to remember state"? maybe "have to remember" is close enough

      Or maybe order is important / do things in the right order is all we need to describe the problem/need.

    1. p. 217:

      We also keep a higher percentage of our assets in cash than most financial advisors would recommend --- something around 20% of the value of our assets outside the value of our house.

      Compare this to my current allotment.

    1. Yes, you do face difficult choices (moral) but you don't care about it. All you care are the reputation bars. So... Let's kill this guy, who cares if he is innocent, but this faction needs it or I'm dead. Sounds great on paper but to be honest... you just sit there and do whatever for these reputation bars. If you won't, then you lose
  5. Jan 2021
    1. I want to write my own scripts first, but may end up graduating to this.

    1. overflow-wrap: break-word; makes sure the long string will wrap and not bust out of the container. You might as well use word-wrap as well because as the spec says, they are literally just alternate names for each other. Some browsers support one and not the other.
    1. Moving DOM elements around made me anxious and I wanted to preserve natural tab order without resorting to setting tabindex, so I also made a flexbox version that never moves DOM elements around. I think it's the superior solution, at least for the layouts I was going for. https://github.com/wickning1/svelte-components/blob/master/src/FlexCardLayout.svelte
    1. Popper for Svelte with actions, no wrapper components or component bindings required! Other Popper libraries for Svelte (including the official @popperjs/svelte library) use a wrapper component that takes the required DOM elements as props. Not only does this require multiple bind:this, you also have to pollute your script tag with multiple DOM references. We can do better with Svelte actions!
    1. I'd like to spin up a couple of these, both for my personal box (localhost-only) and for the development network.

  6. Dec 2020
    1. Add this to my toolchain (in particular, configure Lighthouse to run in our CI/CD pipeline).

    1. I guess it's about "preloading" and not "navigation", if it's the case, then I guess there is still no way to attach to navigation events, and this issue should be kept open.
    2. No JS event is fired, so there currently isn't any clean way to do this that I can see.
    1. You can afford to make a proper PR to upstream.
    2. No more waiting around for pull requests to be merged and published. No more forking repos just to fix that one tiny thing preventing your app from working.

      This could be both good and bad.

      potential downside: If people only fix things locally, then they may be less inclined/likely to actually/also submit a merge request, and therefore it may be less likely that this actually (ever) gets fixed upstream. Which is kind of ironic, considering the stated goal "No more waiting around for pull requests to be merged and published." But if this obviates the need to create a pull request (does it), then this could backfire / work against that goal.

      Requiring someone to fork a repo and push up a fix commit -- although a little extra work compared to just fixing locally -- is actually a good thing overall, for the community/ecosystem.

      Ah, good, I see they touched on some of these points in the sections:

      • Benefits of patching over forking
      • When to fork instead
    1. Some devs prefer Svelte’s minimal approach that defers problems to userland, encouraging more innovation, choice, and fragmentation, and other devs prefer a more fully integrated toolkit with a well-supported happy path.

      tag?: what scope of provided features / recommended happy path is needed?

    2. It’s worth mentioning that Svelte limits its scope to being only a UI component framework. Like React, it provides the view layer, but it has more batteries included with its component-scoped CSS and extensible stores for state management. Others like Angular and Vue provide a more all-in-one solution with official routers, opinionated state management, CLIs, and more. Sapper is Svelte’s official app framework that adds routing, server-side rendering, code splitting, and some other essential app features, but it has no opinions about state management and beyond. Some devs prefer Svelte’s minimal approach that defers problems to userland, encouraging more innovation, choice, and fragmentation, and other devs prefer a more fully integrated toolkit with a well-supported happy path.

      tag?: what scope of provided features / recommended happy path is needed?

    3. It’s worth mentioning that Svelte limits its scope to being only a UI component framework. Like React, it provides the view layer, but it has more batteries included with its component-scoped CSS and extensible stores for state management.
    1. p. 198:

      Given any five points on a sphere, show that some four of them lie on a hemisphere that includes its boundary.

      I'll admit, I already looked at the hint for this problem, and yes, my initial approach did indeed consist of trying to find the 'worst' configuration.

      I can think of two ways to determine whether or not two points on a sphere lie within the same hemisphere:

      • First off, since any two points on a sphere may be connected by a great circle, they're in the same hemisphere if they're separated by no more than \(\frac{\tau}{2}\) radians along this shortest path.
      • Equivalently, the length of the line segment connecting them must be less than or equal to \(2r\), where \(r\) is the radius of the sphere.

      One other note:

      • It's always possible to divide the sphere in half in such a way that any two points lie within the same hemisphere. (This is a corollary of the first point, above. Note that two antipodal points must necessarily fall on the boundary of such a division.)

      So, I have a picture in my mind of the sphere divided into eight regions of equal area by way of three great circles which intersect one another at right angles. (Think the Equator, the Prime Meridian, and a third great circle drawn through the poles at 90 degrees longitude.) My thinking now tends more toward combinatorics and the pigeonhole principle than geometry proper.

  7. Nov 2020
    1. I'd love to take this for a spin. Maybe I could rewrite Demeter or micdrop using it.