triceratops-show / www.triceratops.show Goto Github PK

we can look at the content-length header

< content-length: 147706662

and check for it in a preSave event

CMS.registerEventListener({
  name: 'preSave',
  handler: ({ entry }) => {
    return entry.get('data').set('title', 'new title');
  },
});

decaporg/decap-cms#5306

i guess easiest way is to migrate to static cms?
https://www.staticcms.org/docs/netlify-cms-migration-guide

right now we have pagination (eg docs/page/1, docs/page/2)
i think they come from the .Paginateuse

so i guess we have to remove this query
https://github.com/triceratops-show/triceratops-show.github.io/blob/main/layouts/partials/row.html#L22

luiz asked for it

https://www.netlifycms.org/docs/customization/

Idea 1:

We could try to autopopulate a field

CMS.registerEventListener({
  name: "preSave",
  handler: ({ entry }) => {
    const image = entry.get("data").get("episode_image");

    return entry.get("data").set("image", image);
  },
});

Problem is that it's infered to be a type: image, so we can't just hide it.
https://github.com/netlify/netlify-cms/blob/460d1e6835af122f86278b4b2f9c2e3b7bc91488/packages/netlify-cms-core/src/constants/fieldInference.tsx#L63

Tecnically we could inject some custom css to hide that field

Idea 2:
Rename podcast_image to simply image

It's good to have an idea.

Maybe we should run a cron that access all episodes?

eh right now they are being duplicated after build (to /docs dir)

i guess when moving to a proper server this won't happen anymore

we could pull from our bucket all the static data (mp3 episodes) which we can then refer via hugo, guaranteeing they exist

also we maybe could generate certain fields like episode_bytes and duration based on that?

https://support.deezer.com/hc/en-gb/articles/115000960389-How-to-get-your-podcast-on-Deezer
https://www.facebook.com/luiz.terra1/posts/4665277816838693?comment_id=4665425286823946&__cft__[0]=AZXZS8g-N3HlwqKj0FJ9QAHSutIXjbMOTus7bTmbG--lp3wWBqh-73u7pVe98frRnUTWeS1q2Ttzk1PN7rtjoFpYx2bWaGMSgbbQ4KcYOXRHYQTkAfShn6Qq9IXJJSOOx8c&__tn__=R]-R

https://stackoverflow.com/questions/10622179/how-to-find-identify-large-commits-in-git-history

We could listen to events to autogenerate certain fields:

• podcast_duration
• podcast_bytes

based on podcast_file, which we assume should exist.

In the future we could even do the same for youtube.

https://github.com/porjo/youtubeuploader

step 1 is to run this locally
step 2 is to run in ci, the challenge is how to integrate with netlify cms and how to idempotently update the same video for the same episode

decaporg/decap-cms#5013

since most episodes' body are similar

7 minutes is a lot

every new post shouldn't change much, except the n-1 episode
therefore the deployment shouldn't be so slow (it's taking ~5 minutes currently)

https://developer.mozilla.org/en-US/docs/Web/API/HTMLImageElement/decoding

https://github.com/youtube/api-samples/blob/master/go/upload_video.go

We could create a cli that uploads to youtube
it would save me a few seconds

final objective is to have something that takes the mp3 + the cover art, generates a video and uploads to youtube

For some reason we also build xml files for the artists and participants page

which is around 270 files currently

❯ find public -name '*.xml' -type f | grep -E '(artistas|particip)' | wc -l
     270

eg
https://www.triceratops.show/artistas/13th-floor-elevators/index.xml

so that we don't accidentaly publish episodes

It's just easier to manage everything in the same place.

Now that we display covers under the episodes (#72), maybe we should allow people zooming into these covers, since sometimes they are so detailed (eg https://www.triceratops.show/episodios/43/)

ex: that 2 episodes don't have the same URL

every time i release a new episode i have to hard refresh the index page to see the new episode

i think we have to cache bust that somehow

there are only 21 episodes, last one is from december
https://podcasts.apple.com/us/podcast/tricer%C3%A1tops-show/id1574767544

one really important thing is parsing the xml and validating it's still good, valid rss xml

the website itself could also be tested, ideally with the same tool as we are validating the xml
maybe cypress? :)

ye

https://pfischbeck.de/en/posts/multilingual-dates-in-hugo/

https://gohugo.io/variables/git/