feat(develop): add explicit express handler for page-data requests #27882
Conversation
gatsbot
bot
added
the
status: triage needed
pieh
added
topic: resource loading*
and removed
status: triage needed
| const pageData = await readPageData( | ||
| path.join(store.getState().program.directory, `public`), | ||
| page.path | ||
| ) |
There was a problem hiding this comment.
Just to clarify - we will actually use the new page data utility here? (i.e. #27939)
There was a problem hiding this comment.
That's the plan. It should look similar as https://github.com/gatsbyjs/gatsby/pull/27939/files#diff-e4535fc084e00493cdf64059b0257b3ba2ee3fe54ec9396daf87158b0679905cL193-R200 part in websocket manager
…d dont rely on express-static
new express handler will try to load the page result if page exists in store so to mimic 404 page not existing we actually need to delete that page
pieh
force-pushed
the
dev-page-data-express-handler
branch
from
78d3dba
to
ee81ad6
Compare
| emitter.on(`DELETE_PAGE`, action => { | ||
| if (page404 && action.payload.path === page404.path) { | ||
| boundActionCreators.deletePage({ | ||
| ...page404, | ||
| path: PROD_404_PAGE_PATH, | ||
| }) | ||
| page404 = null | ||
| } | ||
| }) |
There was a problem hiding this comment.
prod-404internal plugin is actually also used ingatsby develop- removing
/404/page (that this plugin clones into/404.html) wasn't removing clone which meant that 404 page removals didn't actually removed/404.htmlpage
There was a problem hiding this comment.
I don't know much about this plugin, just want to understand: why does it hook into onCreatePage instead of creating its own page? And where do we actually create the source 404 page?
There was a problem hiding this comment.
Ok, so explanation for this plugin - when user creates src/pages/404.js, the actual path for page is /404/ (that's what gatsby-plugin-page-creator does) which translated to /404/index.html when generating .html file - this is fine, but a lot of hosting solutions handle /404.html as potential default to show 404 page (I think Netlify does this for sure). So we do special case 404 kind of page and create "duplicate" page for it
The src/pages/404.js is also not only case when we do this - if user creates page programatically in createPages - this will work too.
Note that this is mostly needed for production builds, but our shared runtime (both dev and prod) also uses /404.html in multiple places, so the way this plugin works impact both develop and prod.
pieh
changed the title
| emitter.on(`DELETE_PAGE`, action => { | ||
| if (page404 && action.payload.path === page404.path) { | ||
| boundActionCreators.deletePage({ | ||
| ...page404, | ||
| path: PROD_404_PAGE_PATH, | ||
| }) | ||
| page404 = null | ||
| } | ||
| }) |
There was a problem hiding this comment.
I don't know much about this plugin, just want to understand: why does it hook into onCreatePage instead of creating its own page? And where do we actually create the source 404 page?
Description
Adds explicit handler and not use
express-staticforpage-datafetch requests. This allow us to find exact page browser request resources for and will enable triggering query on demand.Side-effect will also be that we won't serve "stale" page-data resources in develop (builds already handle it by deleting files - #26937 )
[ch18448]