|
Let's be clear: it's not as if git is trying and failing to hide the DAG behind some abstraction: it would be foolish to try to do so precisely because it would be inevitably leaky. There may be other leaky abstractions git presents, but that's not one of them. The point behind hammering importance of understanding that git is modeled as a dag is that it is the core model of the system. Could there be a far better interface for operating on it? Absolutely. Is it a failing of git that understanding that it's operating on a dag is so important? Not in my opinion, since I don't see a clear way personally to design a better system that would successfully abstract over that.
|
# ? Oct 7, 2019 05:31 |
|
|
# ? May 25, 2024 07:28 |
|
One failing of the git interface is that it never actually makes the nature of the graph apparent to the novice user. Perhaps making the ascii-art commit graph the default format of git log would help.
|
# ? Oct 7, 2019 06:08 |
|
There are many improvements that could be made to the concept of a DAG-based version control tool. Git is a good early effort. You could probably get halfway to a next-generation design by setting up a bunch of well-named aliases.
|
# ? Oct 7, 2019 06:58 |
|
Jabor posted:One failing of the git interface is that it never actually makes the nature of the graph apparent to the novice user. It's sort of baffling that the graph isn't on by default. How are you supposed to read the history of anything involving non-ff merges without it?
|
# ? Oct 7, 2019 07:17 |
|
I have no problem understanding DAGs, and I hate git's interface. Normally I know exactly what I want the DAG to look like in the end but it's a right pain to actually get it that way. I had no trouble learning mercurial (similar internal model) or darcs (also decentralised but different, also lol)
|
# ? Oct 7, 2019 07:25 |
Hughlander posted:The only intuitive user interface is the nipple. After that it’s all learned behavior. Spoken like someone who's never had to teach a baby to loving just drink already.
|
|
# ? Oct 7, 2019 09:40 |
|
"Falsehoods programmers believe about user interfaces"
|
# ? Oct 7, 2019 10:14 |
|
You're right. Dags are essential complexity at least for my workflow, and I think of using git as manipulating a dag (but using magit instead of the horrible CLI).
|
# ? Oct 7, 2019 13:53 |
|
That AV has aged badly
|
# ? Oct 7, 2019 16:30 |
|
Xarn posted:That AV has aged badly Much like his foot cheese, and unlike his pal Minsky's preference, it will only get better with age.
|
# ? Oct 7, 2019 17:27 |
|
Xarn posted:That AV has aged badly
|
# ? Oct 7, 2019 17:34 |
|
Twitter killed Javascript. I wishMr.Radar posted:Twitter permanently suspended the official account of the Javascript standards committee for no discernible reason:
|
# ? Oct 9, 2019 03:50 |
|
Javascript has a standards committee?
|
# ? Oct 9, 2019 06:26 |
|
Carbon dioxide posted:Javascript has a standards committee? Not any more
|
# ? Oct 9, 2019 07:11 |
|
He says the email was inactive. I'm guessing he tweets through the API or in some other way triggered a bit flag, didn't respond to a prompt and got wiped. Might as well start a new account with a better profile name. At least it's better than other standards committees who are probably still having greybeard grudge matches on newsgroups.
|
# ? Oct 9, 2019 08:02 |
|
https://twitter.com/EWGchair https://twitter.com/EWGchair/status/1158460632787632128 https://twitter.com/EWGchair/status/1154799581542023169
|
# ? Oct 9, 2019 09:00 |
|
Ola posted:He says the email was inactive. I'm guessing he tweets through the API or in some other way triggered a bit flag, didn't respond to a prompt and got wiped. The email notifying them of the ban (shown in Absurd Alhazred's post) is much more curt and final in its wording than I would expect if it were merely a matter of "hey you're not using this API correctly, change your behaviour please". The email is basically "you're a shithead and you're banned, gently caress off"
|
# ? Oct 9, 2019 09:11 |
|
Hammerite posted:The email notifying them of the ban (shown in Absurd Alhazred's post) is much more curt and final in its wording than I would expect if it were merely a matter of "hey you're not using this API correctly, change your behaviour please". The email is basically "you're a shithead and you're banned, gently caress off" I meant to say "bot flag", as in they thought he was a bot. They probably ban millions of accounts every day and deal with thousands of appealing assholes. So the messages about this are not very polite. His main account is obviously fine so it's not some nefarious personal grudge.
|
# ? Oct 9, 2019 10:21 |
|
Ola posted:He says the email was inactive. I'm guessing he tweets through the API or in some other way triggered a bit flag, didn't respond to a prompt and got wiped. Ah, the long arguments about whether Boost should just move over to CMake.
|
# ? Oct 9, 2019 14:03 |
|
I just had the displeasure of using an online registration form that had the most galaxy-brained password entry field I’ve seen. After you input your password, it would erase it from the form (I assume it also stored it in a cookie or something) and then filled the field with literal * characters. I noticed it was behaving kind of weird (after unfocusing the field a load spinner would appear and then the number of * would change) but I didn’t fully realize what was happening until I saw my password manager had literally stored “******************************” as the password.
|
# ? Oct 9, 2019 15:42 |
|
tankadillo posted:I just had the displeasure of using an online registration form that had the most galaxy-brained password entry field I’ve seen. After you input your password, it would erase it from the form (I assume it also stored it in a cookie or something) and then filled the field with literal * characters. This sounds like Lotus Notes, but Very Online. It had hieroglyphs that would change with every key press when entering a password. It's been 19 years since I used it and I still can't get over how bad it was. boo_radley fucked around with this message at 16:58 on Oct 9, 2019 |
# ? Oct 9, 2019 16:26 |
|
tankadillo posted:I just had the displeasure of using an online registration form that had the most galaxy-brained password entry field I’ve seen. After you input your password, it would erase it from the form (I assume it also stored it in a cookie or something) and then filled the field with literal * characters. That is weird and silly but why aren't you getting the password manager to generate the password for you?
|
# ? Oct 9, 2019 17:24 |
|
Hammerite posted:That is weird and silly but why aren't you getting the password manager to generate the password for you? A common workflow at least with last pass is to generate the pay with the app but then only save it on form submit when it can also get the login and site url. You could go through and create a whole login but may need to guess then what field it will consider the login.
|
# ? Oct 9, 2019 17:28 |
|
Hammerite posted:That is weird and silly but why aren't you getting the password manager to generate the password for you? Sounds like if the manager fills in the field and then the blur even fires for any reason, the value gets replaced. The managers I've used will dutifully record that new value on submit.
|
# ? Oct 9, 2019 17:30 |
|
Hammerite posted:That is weird and silly but why aren't you getting the password manager to generate the password for you? Hughlander posted:A common workflow at least with last pass is to generate the pay with the app but then only save it on form submit when it can also get the login and site url. You could go through and create a whole login but may need to guess then what field it will consider the login. Munkeymon posted:Sounds like if the manager fills in the field and then the blur even fires for any reason, the value gets replaced. The managers I've used will dutifully record that new value on submit. Basically this. My workflow (with 1Password) is: generate password -> password autofills to form -> submit form -> save form data in manager. The last step is kind of optional because it will automatically save the passwords you generate anyway, but saving it after the form submits also saves your username and other form data. Also I often have to generate a bunch of passwords before it finally makes one that conforms to website’s dozen password requirements. tankadillo fucked around with this message at 17:56 on Oct 9, 2019 |
# ? Oct 9, 2019 17:51 |
|
I forgot that a lot of people use password managers that are integrated into the web browser. I just run the password manager application off a usb stick and copy-paste from it.
|
# ? Oct 9, 2019 17:56 |
|
It's not surprising to see the manual pasting and saving of generated passwords flow with LastPass because its browser plugin is hot garbage, at least compared to 1Password and Bitwarden
|
# ? Oct 9, 2019 18:18 |
|
Steve French posted:It's not surprising to see the manual pasting and saving of generated passwords flow with LastPass because its browser plugin is hot garbage, at least compared to 1Password and Bitwarden If you say so? This seems like a non sequitur
|
# ? Oct 9, 2019 18:50 |
|
Hammerite posted:If you say so? This seems like a non sequitur I suppose I was unclear and perhaps unnecessarily inflammatory, then. All three of them, in my experience, have the intended/ideal flow when generating a password that you just have the generator auto-fill the password form, and then the password along with the rest of the login is automatically saved (upon confirmation) for you. In practice, I found this to be a very unreliable mechanism with Lastpass: often it would fail to autofill the correct entry box, or any at all, and more frequently it would fail to correctly capture / collect login details entered across multiple pages, or fail to present the save confirmation after those steps were completed. As a result, I found myself out of habit always generating a password, copying it to my clipboard myself, pasting it into the form, and then proceeding through the login flow so that if/when it failed to properly capture or fill everything, I could more easily correct that. Bitwarden has been much more consistent for me, in that area as well as login autofill, both on desktop browsers and mobile (android) browsers and apps. Especially so on Android. I also saw this the other day, FWIW https://twitter.com/_pastelsky/status/1180864405648502784
|
# ? Oct 9, 2019 19:43 |
|
I’ll be god damned if I’m gonna let my passwords be protected by Javascript
|
# ? Oct 10, 2019 18:59 |
|
Literally Elvis posted:I’ll be god damned if I’m gonna let my passwords be protected by Javascript My bank account and other important stuff are all protected by physical tokens and/or in-memory storage. (Brain memory.) Password managers exist mainly for your werewolf erotica forum accounts and stuff like that. Incidentally, Bitwarden has a lightweight port that runs on Sqlite + Rust and it makes for a great Raspberry Pi installation.
|
# ? Oct 10, 2019 20:13 |
|
Stack Overflow is melting down because they published a new code of conduct that has some stuff in it about pronouns and loads of people caught a case of the "what-ifs". I guess that's inevitable when you have a community that's so obsessed with petty rules-enforcement and process though. also someone got fired or something? It seems like nobody over there is capable of approaching things like an adult for some reason.
|
# ? Oct 10, 2019 20:52 |
|
It's almost as though stackoverflow is mostly populated by bored children screwing around during class or something
|
# ? Oct 10, 2019 21:03 |
|
code:
Oh well, time to rewrite this in a real language with some non-stupid structure!
|
# ? Oct 10, 2019 21:33 |
|
So seeing "addr" and "pe" I thought it was going to be some windows debugging thing until it got talking about city and zip.
|
# ? Oct 10, 2019 22:26 |
|
repiv posted:What's the problem with LunarG? i never responded to this because its a huge rant but thankfully a convenient shorthand has plopped in my lap https://github.com/KhronosGroup/Vulkan-Tools/commit/c60f79013bd8abf65c9b680dfb3958fa27c8c1f6
|
# ? Oct 11, 2019 01:35 |
|
Suspicious Dish posted:i never responded to this because its a huge rant but thankfully a convenient shorthand has plopped in my lap https://github.com/KhronosGroup/Vulkan-Tools/commit/c60f79013bd8abf65c9b680dfb3958fa27c8c1f6 Wow.
|
# ? Oct 11, 2019 02:50 |
|
Suspicious Dish posted:i never responded to this because its a huge rant but thankfully a convenient shorthand has plopped in my lap https://github.com/KhronosGroup/Vulkan-Tools/commit/c60f79013bd8abf65c9b680dfb3958fa27c8c1f6 oh
|
# ? Oct 11, 2019 02:53 |
|
Does mock in the filename indicate that these were just mocks being used for tests or does it mean something else in the context of vulcan? Also super loving bizarre to leave interpreting specs to someone that apparently doesn't understand the basic of the language or how to read a spec doc.
|
# ? Oct 11, 2019 03:04 |
|
|
# ? May 25, 2024 07:28 |
|
Suspicious Dish posted:i never responded to this because its a huge rant but thankfully a convenient shorthand has plopped in my lap https://github.com/KhronosGroup/Vulkan-Tools/commit/c60f79013bd8abf65c9b680dfb3958fa27c8c1f6 Hahahahahahaha
|
# ? Oct 11, 2019 04:38 |