765203 - Support mode "unknown" when parsing cache manifest

Status: RESOLVED FIXED

(Core :: Networking: Cache, enhancement)

Comment 1

[Honza Bambas (:mayhemer)]

We have to allow sections other then e.g. NETWORK, EXPLICIT etc simply by ignoring/skipping them.  This is needed for forward compatibility when the spec adds new sections, as has happened recently.  This should be tracked for FX16 as this is very simple and safe to implement and compat important.

Comment 2

[Honza Bambas (:mayhemer)]

Steeling the bug back.  I would like to get this done for Fx16.

Comment 3

[Honza Bambas (:mayhemer)]

Attachment: v1

- const_iterator doesn't support - operator.. quit sad.
- end cannot be equal to begin since there if it would, the method would exit earlier above (there is if (begin == end || *begin == '#') return NS_OK;)

Comment 4

[Alex Keybl [:akeybl]]

(In reply to Honza Bambas (:mayhemer) from comment #1)
> This should be tracked
> for FX16 as this is very simple and safe to implement and compat important.

The only part of that justification that qualifies for tracking is the compatibility importance. Can you give us more context there? Thanks!

Comment 5

[Jason Duell]

Comment on attachment 641565 [details] [diff] [review]
v1

Both code + test look good.

Comment 6

[Honza Bambas (:mayhemer)]

(In reply to Alex Keybl [:akeybl] from comment #4)
> The only part of that justification that qualifies for tracking is the
> compatibility importance. Can you give us more context there? Thanks!

In the current-work spec it is marked as ready-for-first-implementation.  I checked with chrome (dev channel) that it supports unknown sections.  There has recently been introduced SETTINGS section, that when encountered by our manifest parser leads to break of the manifest load and thus the whole app doesn't get cached.

I'm going to land this patch now, since it is simple, has a test, and I don't see any reason why we should hold it back.

Comment 7

[Honza Bambas (:mayhemer)]

Comment on attachment 641565 [details] [diff] [review]
v1

https://hg.mozilla.org/integration/mozilla-inbound/rev/2ff61044edbd

Comment 8

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/2ff61044edbd