bleedingdeacons / link / 33918537450
92%

DEFAULT BRANCH: main
Ran
Jobs 1
Files 11
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 94.256% (+13.7%) from 80.548%
33918537450

push

github

web-flow 
test: raise Link.Core coverage from 80.5% to 94.3% (#12)

Forty-eight tests, and almost all of them are refusals -- the half that
is never exercised by hand. A message that opens is proved the first
time anybody uses the app; a message that will not open is the one that
reaches a member as an empty screen.

The client's whole password surface had no tests at all: SetPassword,
RequestPasswordLink, and the four small endpoints a handset calls about
itself -- push token, key rotation, key fault, mark read -- were at 0%.
What is asserted throughout is the distinction the client exists to
keep: the server refusing and saying why is not the same as the request
never arriving. Collapsing those is how a member is told their password
is wrong when the wifi is off.

Four of the new assertions were wrong on the first run and two more
passed for the wrong reason: they matched a form-encoded body the client
never emits, so DoesNotContain("code=") would have held with the
credential present. Corrected to the JSON the client actually sends.

The envelope refusals now cover what happens to a message sealed to a
keypair the handset no longer has -- permanent, since Fellowship never
held the private half -- and to one whose ciphertext has been tampered
with, which AES-GCM refuses rather than decrypting to nonsense. Both
must return null rather than throw: they run on the push path, where a
throw kills the process Android started to deliver the message.

The history's recovery paths are covered for the same reason. Reading a
corrupt or foreign-keyed file as empty is a design decision, not a
fallback -- refusing to start because a cache will not decrypt would
take the app away over messages the server can send again.

722 of 766 relevant lines covered (94.26%)

9.77 hits per line

Jobs
ID Job ID Ran Files Coverage
1 33918537450.1 11
94.26
GitHub Action Run
Source Files on build 33918537450