Bug 235490 – [Flatpak SDK] Extension points for the Flatpak Sparkle-CDM extension

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Philippe Normand]

bst bits allowing our Sdk and Platform to have the extension mounted in the runtime sandbox.

[Philippe Normand]

Created attachment 449751 [details]
[fast-cq] Patch

[Adrian Perez]

Comment on attachment 449751 [details]
[fast-cq] Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=449751&action=review

> Tools/buildstream/elements/flatpak/platform.bst:84
> +      no-autodownload: 'true'

What's the reason to quote the booleans? They could as well be specified
directly as they are valid YAML values:

  autodelete: False
  no-autodownload: True

I know that there are many other instances around which are strings, but
the man page for fltpak-metadata clearly states that these fields are
indeed booleans (not strings). Probably this has been working fine because
when saving the data back when building the image the writer that is
serializing the values by chance chooses to write the strings unquoted
(which is allowed in YAML) and the parser from the Flatpak tooling is
magnanimous enough to accept the lowercase unquoted “true“ and “false“
strings as booleans.

Or am I overthinking things here? Anyway, we know the current approach
works and I am mostly wondering what's going on in reality.

[Philippe Normand]

I honestly have no clue what to answer, excepted that iirc I took inspiration in the FDO SDK, maybe the syntax improved since then and this has bitrotten. I can follow-up about this in a separate patch :)

[EWS]

Committed r289118 (246815@main): <https://commits.webkit.org/246815@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 449751 [details].

[Radar WebKit Bug Importer]

<rdar://problem/88491360>