\n\nThat's already on the TODO.
:q!?The reason why I’m asking: I think I sometimes hit
:w in between. This is kind of muscle memory for me (I’m actually not 100% sure if I do it or not …). So a :q! wouldn’t work for me, because I would accidentally send a half-written twt.Another thing to consider: Maybe users have configured their editors to do auto-save after a while. 🤔
Either way, I suspect the current way to cancel a twt (saving an empty file) has to stay (whether I implement an *additional* mechanism to detect “cancel” or not). It is very explicit and the user/editor won’t do it by accident.
So, yeah, what’s your workflow? How do you use it?
:q!?The reason why I’m asking: I think I sometimes hit
:w in between. This is kind of muscle memory for me (I’m actually not 100% sure if I do it or not …). So a :q! wouldn’t work for me, because I would accidentally send a half-written twt.Another thing to consider: Maybe users have configured their editors to do auto-save after a while. 🤔
Either way, I suspect the current way to cancel a twt (saving an empty file) has to stay (whether I implement an *additional* mechanism to detect “cancel” or not). It is very explicit and the user/editor won’t do it by accident.
So, yeah, what’s your workflow? How do you use it?
:q!?The reason why I’m asking: I think I sometimes hit
:w in between. This is kind of muscle memory for me (I’m actually not 100% sure if I do it or not …). So a :q! wouldn’t work for me, because I would accidentally send a half-written twt.Another thing to consider: Maybe users have configured their editors to do auto-save after a while. 🤔
Either way, I suspect the current way to cancel a twt (saving an empty file) has to stay (whether I implement an *additional* mechanism to detect “cancel” or not). It is very explicit and the user/editor won’t do it by accident.
So, yeah, what’s your workflow? How do you use it?
> What’s your workflow here?
I start a twt or a reply/fork, and
:x when done. If I were to change my mind, and wanted to abort, yes, I would :q!. I don’t think you can catch the abort easily, right?
---8<--- Everything below this line will be removed. Empty, save and quit to cancel. ---8<---
--->8--- everything below this line will be removed --->8---
:q!-ing. Or what do you mean exactly by aborting? ggdG clears the whole file, not sure if there's something shorter out of the box.
> what do you mean exactly by aborting?
Aborting is cancelling the twt in progress. Right, that's what I said: deleting all content, saving, and exiting aborts right now.
vim exits with a non-zero exit status on :cq! Hmmm 🤔 Perhaps I’m wrong about that 😁
vim exits with a non-zero exit status on :cq! Hmmm 🤔 Perhaps I’m wrong about that 😁
:cq! does indeed exit with 1. I wasn't aware of this command, though. Cool.
I hit "reply" hotkey and my editor comes up.
With or without writing something I close my editor **without saving the content**.
Of course I close it by C-x C-c, not by :q! ;-)
Jenny finds the temp file unchanged, e.g. it's content is the same as it was when my editor was started. I would like that jenny discards the reply then.
Autosaving is no problem either. Real editors do this to a temporary (kind of backup) file. Only in case of a crash that file is consulted and the user is asked if she would like to continue with that stored content.
I only know this "feature" from my revision control software where I get "abort: empty Commit message" or "Aborting commit due to empty commit message" when I do not change whatever is already in there. Can be quite some text about which files changed and so on.
git rebase -i pre-fills the editor with some commands and when you just exit the editor, Git will execute those commands. If you want to *cancel*, you have to delete all commands and then exit.This commit now adds a little message: http://www.uninformativ.de/git/jenny/commit/9be74280cced4c1e00734f02d2ae98aaedfaff02.html
git rebase -i pre-fills the editor with some commands and when you just exit the editor, Git will execute those commands. If you want to *cancel*, you have to delete all commands and then exit.This commit now adds a little message: http://www.uninformativ.de/git/jenny/commit/9be74280cced4c1e00734f02d2ae98aaedfaff02.html
git rebase -i pre-fills the editor with some commands and when you just exit the editor, Git will execute those commands. If you want to *cancel*, you have to delete all commands and then exit.This commit now adds a little message: http://www.uninformativ.de/git/jenny/commit/9be74280cced4c1e00734f02d2ae98aaedfaff02.html
/dev/brain wasn’t mounted. 🤦
/dev/brain wasn’t mounted. 🤦
/dev/brain wasn’t mounted. 🤦
Writing a twt is more similiar to writing a commit message. Git does quite some checks to detect that nothing **new** was written and happily discards a commit if you just leave the editor. You don't need any special action, just quit your editor. Git will take care for the rest.
But it's OK as it is. I just didn't expect that I have to select and delete all to discard a twt. So it's C-x h C-w C-x C-c for me.
pick-ed then nothing will happen, the commit IDs remain unchanged. However, it reports a successful rebase, which might had thrown us (@movq and me) off. I was also under the impression that even pick-ing rewrites the history. I must have confused this with something else. You just have to use another editor, scnr. ;-D It would be pretty cool to detect an unchanged file, though. But I don't use jenny (yet?), so don't count my vote, @movq. ;-)
Confusion arises from the fact that jenny *pre-fills* parts of a twt when you reply to someone. When you just quit the editor in that situation, then you didn’t write an empty twt (it contains a conversation hash and a mention), so it won’t be discarded.
That’s why I compared it to
git rebase and not git commit. git rebase pre-fills your editor with some commands as well, and to abort all this, you have to delete them, save, quit. (A rebase just *happens* to not do anything if you pick everything, but it does process your commands – they just happen to be all no-ops. It doesn’t just abort and discard.)Okay, fine. jenny isn’t Git, so it doesn’t really matter anyway.
Is the current behaviour good UX? I don’t know. “Delete everything, save, quit” is pretty much hard-wired into *my* brain, so it’s what *I* expect. So it’s good UX for *me*.
It would be easy to implement an
mtime check (so that jenny aborts if you never save the file at all). I already did it in a local branch, it’s a diff that changes three lines. What I don’t like about this approach: It introduces a heuristic that tries to detect if the user changed the file or not. So what if the user does save the file (because they routinely exit their editor with “save-quit” or whatever) but they didn’t change anything? mtime check will fail, so we have to compare the file’s contents now. And now we’re in a situation where some content is *illegal* to have in a twt – you can now no longer write a twt that contains only (#convhash) @<some person>. Is that a problem? Maybe not, but it means you can’t just bump a conversation with a mention (we routinely do stuff like that at work). Also, in general, I’d like users to be able to put anything they like in a twt, and that won’t work anymore now.So … I don’t know. I’m happy with “delete to cancel”, because it is a very explicit command. *And* it’s easy to implement.
If people are truly unhappy with this (I don’t think you are as you said it’s unexpected but okay), then I’d rather implement an interactive prompt after the user has left the editor: “Do you want to publish this twt [Y/n]? ‘$quote_of_twt’” (In fact, my
publish_command does that. It’s not very UNIX-y, but it’s useful. Sometimes, I change my mind about writing a twt a split second after I left the editor. 🥴 Or stuff like that.)
Confusion arises from the fact that jenny *pre-fills* parts of a twt when you reply to someone. When you just quit the editor in that situation, then you didn’t write an empty twt (it contains a conversation hash and a mention), so it won’t be discarded.
That’s why I compared it to
git rebase and not git commit. git rebase pre-fills your editor with some commands as well, and to abort all this, you have to delete them, save, quit. (A rebase just *happens* to not do anything if you pick everything, but it does process your commands – they just happen to be all no-ops. It doesn’t just abort and discard.)Okay, fine. jenny isn’t Git, so it doesn’t really matter anyway.
Is the current behaviour good UX? I don’t know. “Delete everything, save, quit” is pretty much hard-wired into *my* brain, so it’s what *I* expect. So it’s good UX for *me*.
It would be easy to implement an
mtime check (so that jenny aborts if you never save the file at all). I already did it in a local branch, it’s a diff that changes three lines. What I don’t like about this approach: It introduces a heuristic that tries to detect if the user changed the file or not. So what if the user does save the file (because they routinely exit their editor with “save-quit” or whatever) but they didn’t change anything? mtime check will fail, so we have to compare the file’s contents now. And now we’re in a situation where some content is *illegal* to have in a twt – you can now no longer write a twt that contains only (#convhash) @<some person>. Is that a problem? Maybe not, but it means you can’t just bump a conversation with a mention (we routinely do stuff like that at work). Also, in general, I’d like users to be able to put anything they like in a twt, and that won’t work anymore now.So … I don’t know. I’m happy with “delete to cancel”, because it is a very explicit command. *And* it’s easy to implement.
If people are truly unhappy with this (I don’t think you are as you said it’s unexpected but okay), then I’d rather implement an interactive prompt after the user has left the editor: “Do you want to publish this twt [Y/n]? ‘$quote_of_twt’” (In fact, my
publish_command does that. It’s not very UNIX-y, but it’s useful. Sometimes, I change my mind about writing a twt a split second after I left the editor. 🥴 Or stuff like that.)
Confusion arises from the fact that jenny *pre-fills* parts of a twt when you reply to someone. When you just quit the editor in that situation, then you didn’t write an empty twt (it contains a conversation hash and a mention), so it won’t be discarded.
That’s why I compared it to
git rebase and not git commit. git rebase pre-fills your editor with some commands as well, and to abort all this, you have to delete them, save, quit. (A rebase just *happens* to not do anything if you pick everything, but it does process your commands – they just happen to be all no-ops. It doesn’t just abort and discard.)Okay, fine. jenny isn’t Git, so it doesn’t really matter anyway.
Is the current behaviour good UX? I don’t know. “Delete everything, save, quit” is pretty much hard-wired into *my* brain, so it’s what *I* expect. So it’s good UX for *me*.
It would be easy to implement an
mtime check (so that jenny aborts if you never save the file at all). I already did it in a local branch, it’s a diff that changes three lines. What I don’t like about this approach: It introduces a heuristic that tries to detect if the user changed the file or not. So what if the user does save the file (because they routinely exit their editor with “save-quit” or whatever) but they didn’t change anything? mtime check will fail, so we have to compare the file’s contents now. And now we’re in a situation where some content is *illegal* to have in a twt – you can now no longer write a twt that contains only (#convhash) @<some person>. Is that a problem? Maybe not, but it means you can’t just bump a conversation with a mention (we routinely do stuff like that at work). Also, in general, I’d like users to be able to put anything they like in a twt, and that won’t work anymore now.So … I don’t know. I’m happy with “delete to cancel”, because it is a very explicit command. *And* it’s easy to implement.
If people are truly unhappy with this (I don’t think you are as you said it’s unexpected but okay), then I’d rather implement an interactive prompt after the user has left the editor: “Do you want to publish this twt [Y/n]? ‘$quote_of_twt’” (In fact, my
publish_command does that. It’s not very UNIX-y, but it’s useful. Sometimes, I change my mind about writing a twt a split second after I left the editor. 🥴 Or stuff like that.)
There are three ways to cancel now:
1. Save an empty twt.
2. Quit the editor without ever saving anything to the file.
3. Quit the editor with an exit code other than 0 (":cq!" in Vim).
I think this is good enough. I won’t add anything that works on the twt’s content (i.e. “same content = cancel”), because I want to be able to write twts with arbitrary content (even if it happens to be identical to that pre-filled content).
There are three ways to cancel now:
1. Save an empty twt.
2. Quit the editor without ever saving anything to the file.
3. Quit the editor with an exit code other than 0 (":cq!" in Vim).
I think this is good enough. I won’t add anything that works on the twt’s content (i.e. “same content = cancel”), because I want to be able to write twts with arbitrary content (even if it happens to be identical to that pre-filled content).
There are three ways to cancel now:
1. Save an empty twt.
2. Quit the editor without ever saving anything to the file.
3. Quit the editor with an exit code other than 0 (":cq!" in Vim).
I think this is good enough. I won’t add anything that works on the twt’s content (i.e. “same content = cancel”), because I want to be able to write twts with arbitrary content (even if it happens to be identical to that pre-filled content).