FHIR Chat · Firely Terminal: install bakes everything · tooling

Stream: tooling

Topic: Firely Terminal: install bakes everything


view this post on Zulip Patrick Werner (Feb 10 2022 at 12:26):

When using firely terminal (2.5.0-beta-7) install on a file, Firely Terminal seems to bake the whole cache?

fhir install packages/molit.fhir.common-terminology-0.2.0.tgz --file
Added: molit.fhir.common-terminology 0.2.0.
Baked de.basisprofil.r4@1.2.0.
Baked hl7.fhir.r4.core@4.0.1.
Baked de.gematik.isik-basismodul@1.0.3-alpha1.
Baked kbv.mio.ueberleitungsbogen@1.0.0-kommentierung.
Baked kbv.basis@1.2.0.
Baked molit.fhir.common-terminology@0.2.0.
Restore completed.

fyi: @Ward Weistra

view this post on Zulip Oliver Egger (Feb 10 2022 at 12:41):

did not know that what you can bake with :fhir: , but cool name for creating snaphshots :grinning: , is baking hl7.fhir.r4.core@4.0.1. not dangerous?

view this post on Zulip Patrick Werner (Feb 10 2022 at 13:18):

yes the bake command is quite new, install now also bakes by default

view this post on Zulip Patrick Werner (Feb 10 2022 at 13:18):

hl7.fhir.r4.core is already containing snapshots, so yes it is dangerous.

view this post on Zulip Martijn Harthoorn (Feb 10 2022 at 14:24):

Bake keeps track, so it doesn't bake twice. But it does by default (try to) bake all dependencies on a new install

view this post on Zulip John Moehrke (Feb 10 2022 at 14:27):

so, no good for "twice baked potatoes"...

view this post on Zulip Martijn Harthoorn (Feb 10 2022 at 15:35):

We could do a

>  fhir bake --potato-style

view this post on Zulip John Moehrke (Feb 10 2022 at 15:39):

propose that for an April 1st release

view this post on Zulip Chris Moesel (Feb 10 2022 at 15:43):

I think @Patrick Werner is right though. It probably shouldn't overwrite snapshots in packages that are shipped w/ the snapshots already (e.g., packages published by HL7), right?

view this post on Zulip Ward Weistra (Feb 10 2022 at 15:44):

@Patrick Werner Are you sure this is the whole cache and not only packages within the scope (your installed packages, its dependencies and their dependencies)?

view this post on Zulip Ward Weistra (Feb 10 2022 at 15:45):

I'm not sure if there's already a common way to indicate that a package has been baked/has snapshots for all its resources.

view this post on Zulip Patrick Werner (Feb 10 2022 at 16:11):

Ward Weistra said:

I'm not sure if there's already a common way to indicate that a package has been baked/has snapshots for all its resources.

maybe place a .file in the package folder?

view this post on Zulip Chris Moesel (Feb 10 2022 at 17:01):

I would think that usually a package either has snapshots for ALL its resources or NONE. It seems it would be uncommon for a package to be published with a mix of SDs with and without snapshots. If that's a fair assumption, then I would think you can just look at the first SD you find and check if it has a snapshot or not.

view this post on Zulip Ward Weistra (Feb 10 2022 at 17:10):

Or the bake command could just check all files as it does now and ignore the files already having a snapshot. Not sure what the current behavior is. @Martijn Harthoorn

view this post on Zulip Patrick Werner (Feb 10 2022 at 17:49):

Ward Weistra said:

Patrick Werner Are you sure this is the whole cache and not only packages within the scope (your installed packages, its dependencies and their dependencies)?

yes. These aren't sub-dependencies.

view this post on Zulip Ward Weistra (Feb 11 2022 at 12:58):

@Patrick Werner Ok, that's not as intended. We'll reproduce and fix.

view this post on Zulip Ward Weistra (Feb 11 2022 at 12:59):

Also checking whether bake recreates snapshots for resources that already have them. If so I'll see if that can be optional and not recalculate existing snapshots by default.


Last updated: Apr 12 2022 at 19:14 UTC