david.j.smith
2018-04-17 08:48
@mboudreau I'm in the random channel for a random but not private question: Have a seen you lurking around the Haxe community?
mboudreau
2018-04-17 10:14
@david.j.smith yes. Long time ago. I used to do actionscript and flex for SPAs. Do conferences and such
david.j.smith
2018-04-17 10:17
Thought so. :slightly_smiling_face:
I'm ex-Flash too. Was playing with Haxe recently and recalling of the good old days.
mboudreau
2018-04-17 11:21
Yeah. I find it hilarious that all the hubbub about JavaScript frameworks was things I was working on a decade ago. We definitely went a leap backwards in terms of developer experience
mboudreau
2018-04-19 01:55
can't say that actionscript/flex was a dark secret. I was quite proud of the work I did and it was a great language to build SPA
mboudreau
2018-04-19 01:56
if only adobe actually pushed the boundary and created something that built javascript from actionscript, it would have been the typescript of it's time
david.j.smith
2018-04-19 08:57
Indeed, my dark secret is that I actually still love it (in the form of Haxe), but because unfortunately I've been ruined by Functional Programming and can't go back to OO other than for odd jobs. :joy:
Watching Haxe's development with interest though as they've started adding FP constructs. :crossed_fingers:
ashish.dubey91
2018-09-14 11:04
this happened today https://twitter.com/dash1291/status/1040528107818692608
heytaco
2020-08-25 03:57
Hi there! My name is HeyTaco!, and you can use me to give people tacos to show your appreciation. My tacos will spread joy through Slack!
phil.endsley
2020-08-27 00:46
Hey all, this isn't directly related to pact, but I think people in this slack are the target demographic I'm looking for :slightly_smiling_face:
1. Do you think companies should have dedicated test engineers?
2. What's the difference in responsibility between a test engineer and "non-test" engineer?
ian.letourneau
2020-08-29 15:22
Hi Phil! First what do you mean by "test engineers"? Every companies and individuals I worked with seem to have a different definition :slightly_smiling_face:
phil.endsley
2020-08-29 18:17
Good question! And part of why I asked haha whatever it means to you I guess
I'm a big believer that most, if not all, automated tests should be done by every developer. It's just part of the job.
Our company doesn't have any type of "test engineer" position. Testing is also a huge piece that most teams are missing. I'm considering proposing it as a new position to change that, and wanted to get input from others about how they feel about such positions.
In my head, they would be responsible for:
- driving best practices for automation
- develop tooling/libraries that others can consume to make testing easier
- maybe takeover higher level/more complex test suites?
ian.letourneau
2020-08-29 22:10
> Testing is also a huge piece that most teams are missing
What have you (you, your company or your colleagues) tried until now to improve this situation?
I am trying to understand your context a little bit more before giving my opinion :slightly_smiling_face:
phil.endsley
2020-08-29 22:31
Company - nothing, except say "tests are good". In practice, tests are often dismissed and not prioritized, because there are too many bugs in production :face_with_rolling_eyes:
Me and a handful of colleagues:
- Lead by example
- Require tests for all PRs before merging
- Provide examples of how to write different kinds of tests
- Host workshops, lunch and learns, etc...
We basically have a grassroots movement to make testing part of our culture.
ian.letourneau
2020-08-29 22:36
Do most of your team members understand "what's it in for them"? By that I mean do they understand how automating tests could help them in their day to day job?
phil.endsley
2020-08-29 23:19
Other people on my team, yes, because they've seen the light :grin:
Other devs in the company, they say yes, but I don't think they really get it. Most pushback we get is: yeah, tests are good, but:
- the legacy code base is too hard to test
- this change is really straight forward, so it doesn't make sense to waste time with tests
- there's no way there could be a big in this, so tests are pointless
- we haven't ever done them before, why start now
- I'm too busy to write tests
- most tests I've seen are bad ones that don't actually test anything, so I don't bother writing them
ian.letourneau
2020-08-29 23:35
So if I understood correctly, you're looking for help to make testing part of your culture in a context where not everyone see the value of such idea and you want to know if having someone dedicated to the role of "test engineer" would help.
phil.endsley
2020-08-29 23:52
I guess I'll add, I think it would be, but I do welcome input from others.
I was also just curious how others viewed positions like that (and what does that position do at their company)
ian.letourneau
2020-08-30 00:01
Ok then based on what I could grasp of your context (probably just the tip of the iceberg), on my experience and on what I saw with most of my clients (I may be biased, I'll explain why later), here's my current opinion:
ian.letourneau
2020-08-30 00:01
1. It could have some dangerous side effects to have dedicated test engineers as it could reinforce the idea of having very specific responsibilities and that testing is not part of our job. But that risk could me mitigated if it is very clear that test engineers are here to teach first (like you and some of your colleagues seem to do). Once your movement to make testing part of your culture rises, then it could be helpful to have someone with a better expertise to help building tools more adapted to your context. Or to highlight some blindspots you might have in your process to make sure problems are caught early.
ian.letourneau
2020-08-30 00:07
2. I think the difference in responsibility is that test engineers should be very aware of all the things that could go wrong in the whole process of building software and as every experts they should focus on teaching their expertise instead of doing the work by themselves. In contrast, I think "non-test" engineers should be aware of how to build sustainable software (maintainable and evolving). But that also mean writing good automated tests as they are a good way to criticize our architectures and make sure everything work as expected :slightly_smiling_face:
ian.letourneau
2020-08-30 00:14
But just like you, I would be very curious to know what the others think about this!
slack1
2020-08-31 15:03
i am very much of the mindset that the engineer (or pair, etc) that writes a feature is responsible for proving that it works as defined and that regressions are either prevented or at least easy to find/resolve when they happen. there are proven tools and practices available to make this possible with proper investment and self discipline (and of course iteration and evolution over time)
slack1
2020-08-31 15:06
where i think a dedicated tester/team is still valuable, though, is for exploratory testing. someone looking for edge cases where behavior isnt well defined or is defined incorrectly. exploring various browsers and screen sizes for rendering issues, or interaction issues with mouse vs touch. these folks are not responsible for finding regressions in things that used to work (because that is handled by test suites) or verifying that new functionality works as defined (also handled by test suites), but instead identifying things that feed the backlog of improvements for the system
danhitchcock
2020-09-09 13:17
In perhaps your case where you're fighting technical debt, a dedicated qa is going to have impact.
Having devs that know techniques, still means they need the time to write tests and keep them in good shape. Even the most dedicated person can't wear both hats all the time, so they should at least focus on unit tests a priority.
I'd say my answer would be closer to the one directly above me if the project is a green field and you're not playing catch up.
Anyway, my 2 pence worth.
sklein
2020-10-07 20:37
question... how is the http://pact.io documentation site built? It looks like there's something that aggregates all those readme.md files across the various repos; if so, how?
tjones
2020-10-07 23:04
There's a github workflow thing in each repo - here's the one for pact-js https://github.com/pact-foundation/pact-js/blob/master/.github/workflows/trigger_pact_docs_update.yml
johnnycareer
2021-04-06 21:47
Hello! I am working with a team that would like to start contract testing via Pact within their Azure DevOps CI/CD pipelines environment.
They will run with it long-term. But up front now, they are very busy. And so, I would like to set up Pact within Azure DevOps as much as possible before interrupting them from their existing work.
I am new to Pact and Contract Testing, but have read through the existing Pact and Pactflow materials quite a bit. Was a software engineer for many years, but now very stale. So looking for the best entry point to get this going.
tjones
2021-04-08 03:43
You'll get a better response in #general , but the best thing you can do would be to set up a broker and get an example going. There's a good documentation page for this, hold on
johnnycareer
2021-04-08 21:26
Yes, I was hoping to follow that. So, glad to get a confirming opinion. But too much technical stuff for me to handle alone (at this stage of staleness). I am working with the team to get them involved sooner than later. Thanks!
warren
2021-10-13 16:59
I'm not sure where the best channel for asking general pact questions is, but I've got a sort of conceptual problem. I'm using pact-js.
I've got an asyncronous service that publishes three types of updates to a channel: "pending", "complete", "failure". This service has two code paths that publish these updates. The first is the "initial job submit" codepath, which will publish either a "pending" or a "failure" response. Secondly there is a "job finished" codepath that can publish "complete" or "failure" messages. Because this is an async protocol, the consumer doesn't know which code path actually published the update.
So when testing this service, the "submitJobTests" MessageProvider can only generate pending and failure responses while the "finishJobTests" MessageProvider only generates failure and complete responses. So both tests fail verification because they omit one of the message types. If I "unioned" these tests together I could completely fulfill the pact, but then there are two possible paths for generating "failure" messages and I could only verify one.
Is my approach incorrect? Perhaps I should be putting the emitters into separate utility functions and just testing those? Or is there some way to say this provider does not provide certain types of responses?
tjones
2021-10-14 23:21
Pragmatically, I would solve this with two (from pact's perspective) consumers, which have pacts with the different providers
tjones
2021-10-14 23:22
one that listens for "complete" + "failure", and one that listens for "pending" and "failure"
tjones
2021-10-15 00:41
In JS, you could probably write a super lightweight DSL that let you say "hey I have a pact with these providers X and Y, I'm expecting the following messages to be generated by X +Y, and this message to be only generated by Y"
tjones
2021-10-15 00:41
This is probably a common case for messages, so it might be worth us publishing it in `jest-pact` or something
warren
2021-10-15 20:54
I ended up putting the emitter routines into a seperate utility functions, so then I just use pact to verify those output the expected methods. This seemed to work okay.
tjones
2022-08-12 05:44
I have wanted to have a channel to ask off-topic tech questions (eg API design). #general didn't feel right, and neither did #random
tjones
2022-08-12 05:44
also I don't have whatever question I wanted to ask any more. I can't remember what I did. Probably asked @abubics
yousafn
2022-09-09 10:53
I also wanted this, because I wanted to ask about advice on mono-repos because Lerna was unmaintained, and probably overkill for what I needed.
Something like
general-tech
dev-chat
> A place for more general conversation about technology not necessarily related to Pact
yousafn
2022-09-09 10:56
> think #random is ok :facing_think:
It probably is, but just limited audience compared to some of the other channels, and people may not want to add random for additional noise, if they are just here for the tech related chatter or help.
Mind you a new channel needs to be seeded with people first and takes ages. Probably just overthinking and should just post in general :sweat_smile:
cfmack
2022-11-27 15:58
The Optimist says the glass is half full. The Pessimist says the glass is half empty. The Engineer asks why the glass was built twice the size. -Malcolm Gladwell
yousafn
2022-11-30 19:03
Love that @cfmack :sweat_smile:
Also yay to life in random! :smiling_face_with_3_hearts:
yousafn
2022-12-06 12:03
Hope every one has opened their advents calendars and looking forward to crimbo.
It would be about this time, when I worked back in an office job rather than the open source, that we would be gearing up for change freezes, and smashing in loads of releases before the office becomes a ghost town.
The joys of software delivery :sweat_smile:
matt.fellows
2022-12-06 20:43
Who else was always successful (or maybe not), in getting approval to deploy during such freezes?
matt.fellows
2022-12-06 20:44
I recall at least 3 years running a xmas eve (or last working day before xmas) release :laughing:
yousafn
2023-01-18 18:04
Very impressed with UTM
https://docs.getutm.app/
Have been able to get ARM copies of Windows/MacOS & Linux running on my M1 Mac.
Their documentation is particularly nice, I love their labelling system.
I've used many of the other virtualization tools on a mac (Fusion/Parallels/VirtualBox) but QEMU via UTM is pretty decent so far.
Also for anyone following the efforts for getting Linux installed on an M1 mac, the Asahi Project has been doing incredible things
https://asahilinux.org/
and it has come on leaps and bounds since I last looked, managed to get a copy of Ubuntu running on my M1 - https://github.com/tobhe/ubuntu-asahi
Not done much with it so far, bar getting it installed but looks promising.
matt.fellows
2023-01-19 02:01
Is the idea of Asahi to replace your OSX install with Linux/dual boot? Or more like a virtual machine?
yousafn
2023-01-25 14:12
Asahi allows for dual booting, you can run linux ARM vm's natively in MacOS via QEMU/UTM or Parallels. (I've got a copy of Parallels on my home machine, it is VERY nice, you get app's from the VM's in your host macs dock)
Dual Booted, you still need the mac os system, basically asahi pretends it's a macos install at startup to allow it semi bypass security conditions.
It's not quite ready as a full replacement, no Thunderbolt 4, no display port / hdmi and no sound on a macbook, so I can't get an external monitor.
Nice to play around with, and it is bloody quick, like 1/3 quicker compiling 26 projects (my hello_ffi project)
need to test it with Docker, as that should be alot quicker, and macos is a bit of a hack for Docker.
As I've got an m1 machine for work and m1 personal laptop, I planned to just have it running linux and could vnc or ssh into it when needed.
I have an i9 hackintosh and windows 11 pc which sit connected to the network too, mainly with the aim of connecting them up with the iPad so I could use it as a decent workstation.
Other adventures have been running UTM on my Ipad, running code-server on it, port forwarding the connection, and running it with a safari PWA.
I'm setting the same up for rasp pi, but UTM is working quite nicely, also it is in emulation mode (which is a bit slow, ipad allows for JIT emulation but its not able to take advantage of the m1's speed)
matt.fellows
2023-01-27 00:03
wow, that speed improvement is? incredible. One would have thought the apple engineers would have optimised it. I guess Asahi probably removes a bunch of the non essential things that Apple has added for UX/experience?
yousafn
2023-09-04 15:22
Some updates in the land of MacOS Virtualisation -
If you have a macos, and you want more than 2 virtual machines using apples new virtualisation framework, there is now a way. :rocket:
https://khronokernel.github.io/macos/2023/08/08/AS-VM.html
yousafn
2023-09-04 15:23
Also Asahi Linux is switching up to Fedora for their flagship distro
https://asahilinux.org/2023/08/fedora-asahi-remix/
