385040 results for ""a""

• I could build an API on my site that handles it, maybe with a backing DB or just a JSON file on R2 (S3 equivalent) but then for my deployment to succeed the existing site must be working which is a strange issue.
  autumnlilybug at 2025-12-22 05:13
• hmmmm, you could theoretically stash the data as a json file on your site
  ulhar4409 at 2025-12-22 05:13
• All in all I don't think it's practical for me to automate it in my own system, not to say that automation wouldn't be beneficial, but the effort to make a functional system is too high compared to simplifying the process of manually sending webmentions (which is the point of GET-style webmention)
  autumnlilybug at 2025-12-22 05:12
• Really the issue is state management. A GH action can store the state in the repo, but it can't guarantee that invalid webmentions are never sent. A hook into the deployment process on Cloudflare can guarantee that only valid webmentions are sent but must have an outside system to maintain state which it has to both read from and write to
  autumnlilybug at 2025-12-22 05:11
• but you know your system better than i do so it might be a non-problem
  ulhar4409 at 2025-12-22 05:09
• But yeah, a git precommit hook is imo not a good solution. GitHub actions *could* work but if the Cloudflare deployment fails and the action does not I've sent out invalid webmentions which is not great. So the yarn step is the only reliable way to handle it.
  autumnlilybug at 2025-12-22 05:07
• ulhar4409[d]: For my site, if I commit to the remote repo Cloudflare automatically rebuilds and redeploys from that commit. Adding a new post just means writing an MD(X) file in the correct directory, commit+push, then wait until it redeploys. The only hooks available are git precommit hooks, GitHub actions and the custom `yarn cibuild` action run by the Cloudflare deployment process
  autumnlilybug at 2025-12-22 05:03
• URL is probably a reasonable ID as far as the state data goes
  ulhar4409 at 2025-12-22 05:03
• Then, at least for my setup, it has to read every post and look at the frontmatter (yaml block at the start of the markdown encased in ---). If that post's ID is in the list, skip. Otherwise, if it is marked as published, figure out it's URL and send a webmention to every link in the post (involves discovering endpoints for various hosts - could be cached??), then add to the list
  autumnlilybug at 2025-12-22 05:01
• Oh...I was thinking it was more 'abstract' a meaning
  GWG at 2025-12-22 05:00

Filter results by:

• indieweb (185242)
• dev (98705)
• meta (40044)
• microformats (21687)
• wordpress (18863)
• social (13104)
• known (3994)
• stream (3211)
• bridgy (162)
• events (26)

Query took 0.03s.

Search tips

Exclusion

+foo -bar

Logical OR

foo OR bar

Exact phrase

"foo bar"

Partial words

foo*

Particular fields only

title:foo

domain:example.org

nick:somebody

after:2016-11-23

before:2016-11-23

date:2016-11-23