Closed StefanKarpinski closed 3 years ago
Merging #107 (427b657) into master (ada4f12) will not change coverage. The diff coverage is
100.00%
.
@@ Coverage Diff @@
## master #107 +/- ##
=======================================
Coverage 96.30% 96.30%
=======================================
Files 4 4
Lines 649 649
=======================================
Hits 625 625
Misses 24 24
Impacted Files | Coverage Δ | |
---|---|---|
src/Tar.jl | 96.10% <ø> (ø) |
|
src/create.jl | 97.09% <100.00%> (ø) |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update ada4f12...427b657. Read the comment docs.
While implementing support for hardlinks, the truncated tarball test was causing problems because it was simply cutting the test tarball in half, which is neither guaranteed to create an ill-formed truncated tarball at all (it is easily possible for half a tarball to be a valid tarball by itself), nor guaranteed to consistently truncate the tarball in the same way on different platforms since we generate somewhat different test tarballs on different platforms (depending on: whether we are able to create hardlinks and what the max file name length is). The fact that this test worked previously was essentially dumb luck.
This change replaces the previous test with systematic tests that various specific truncations of a simple, consistent tarball trigger EOFErrors as expected in all the API functions that consume tarballs. It turns out that several of these functions do not error as expected in all circumstances.