297181 - Extension manager fails to identify the default theme at startup

Status: RESOLVED DUPLICATE of bug 295855

(Toolkit :: Add-ons Manager, defect)

Description

[Daniel Brooks [:db48x]]

At startup the extension manager goes through and looks for extensions and
themes that have been placed in the profile or the program folder but not
actually installed yet, and installs them. In the process it finds the default
theme and attempts to install it. However, because the default theme is special,
and we know it'll always be compatible with the browser, the extension manager
attempts to skip the actual install and merely add it to the list of properly
installed extensions. However, this check fails, and it procedes with the
installation. Unfortunately, because the theme is special (and therefore it
seems that nobody had kept it's rdf file up to date), it fails the version
check, causing it to appear to be incompatible with the browser. As a result,
the default theme is never added to the list of installed extensions, and the
same thing happens on every startup.

This isn't a major problem; not having the default theme officially installed
doesn't seem to cause any real problems, but it ought to be corrected.

See also bug 287038, which is about the hardcoded uuid in the extension manager.
This uuid is how the extension manager attempts to identify the default theme.
It's probably possible to fix that bug while still letting the extension manager
special case the default theme by it's id, by using the preprocessor to include
that information at build time. That would allow a derived product to use a
different default theme without making them edit this part of the code as well.

This isn't a recent regression, I think it happened around the time of the last
major extension manager checkin. I've been meaning file this since then, but I
guess I've been busy.

Comment 1

[Robert Strong (they/them - no direct email)]

I highly suspect this was fixed by bug 295855 but that depends on the build you
were using.

Comment 2

[Robert Strong (they/them - no direct email)]

In reading comment #0 again I am quite sure this is a dupe of bug 295855 except
in respect to the comments about bug 287038.

*** This bug has been marked as a duplicate of 295855 ***