4 ms·

I accidentally pasted a backtrace into a github issue outside of a code block. It was a deep stack, with each line prefixed by "#0" "#1", "#2", etc. GitHub deci

I accidentally pasted a backtrace into a github issue outside of a code block. It was a deep stack, with each line prefixed by "#0" "#1", "#2", etc. GitHub decided to turn these into links to the corresponding numbered issues, for every issue, (including back-reference notifications in each of those issues) and there didn't seem to be any way to undo that action.

dwaltrip 3y ago
Oof. There should definitely be an undo button for that, which displays right afterwards.
pm215 3y ago
For QEMU's gitlab issue tracker we deliberately started our issue numbering from a higher number (I think 100 or thereabouts) to avoid that. (We ran into it very early on during the migration from launchpad so it wasn't a big deal to re-number the two or three issues that had already been created with low numbers.)
veave 3y ago
>there didn't seem to be any way to undo that action. Well thank God there isn't a button to undo sent mails.
- jkaplowitz 3y ago
  In many email systems there is an undo send feature. Outside of those closed contexts where yanking a message out of the recipient’s inbox is possible, it is generally implemented by waiting a few seconds between when the UI says the message has been sent and when the message is actually sent. During that interval, an undo send option is offered which simply cancels the upcoming send action.
  - onionisafruit 3y ago
    Nothing calls attention to an email better than those messages saying that the sender wishes to retract a message you have already downloaded.
    - jkaplowitz 3y ago
      Yes, and those recall messages only exist in the closed systems where such retraction is possible, or in the open systems which send those messages that are more designed for the closed systems. The implementation I described with a slightly delayed send doesn’t use those recall messages, since the email being sent doesn’t actually get sent until after the delay. I believe this is how Gmail’s version of the feature works, for example.
moffkalast 3y ago
Sounds like you should... open an issue.
remram 3y ago
A bigger problem I've had with issue references is in commits. A repo we used as template was using "Fixes #123" in commits messages, to close tickets filled against that template repo. The problem is that when the updated template was merged in the downstream repo, it would close issues in each of them too...

dwaltrip 3y ago

Oof. There should definitely be an undo button for that, which displays right afterwards.

pm215 3y ago

For QEMU's gitlab issue tracker we deliberately started our issue numbering from a higher number (I think 100 or thereabouts) to avoid that. (We ran into it very early on during the migration from launchpad so it wasn't a big deal to re-number the two or three issues that had already been created with low numbers.)

veave 3y ago

>there didn't seem to be any way to undo that action. Well thank God there isn't a button to undo sent mails.

jkaplowitz 3y ago
In many email systems there is an undo send feature. Outside of those closed contexts where yanking a message out of the recipient’s inbox is possible, it is generally implemented by waiting a few seconds between when the UI says the message has been sent and when the message is actually sent. During that interval, an undo send option is offered which simply cancels the upcoming send action.
- onionisafruit 3y ago
  Nothing calls attention to an email better than those messages saying that the sender wishes to retract a message you have already downloaded.
  - jkaplowitz 3y ago
    Yes, and those recall messages only exist in the closed systems where such retraction is possible, or in the open systems which send those messages that are more designed for the closed systems. The implementation I described with a slightly delayed send doesn’t use those recall messages, since the email being sent doesn’t actually get sent until after the delay. I believe this is how Gmail’s version of the feature works, for example.

moffkalast 3y ago

Sounds like you should... open an issue.

remram 3y ago

A bigger problem I've had with issue references is in commits. A repo we used as template was using "Fixes #123" in commits messages, to close tickets filled against that template repo. The problem is that when the updated template was merged in the downstream repo, it would close issues in each of them too...