1. Rename various shorthand variables
2. Split simple tests into separate Google Test cases (have not tackled
the more complex tests which have loops and/or setup)
3. Remove redundant logging from simple tests
Relates-to: #15913
Consolidate test frameworks and cleanup, including:
* all tests use Google Test
* reorganize folders, esp. kmnkbd -> api
* move all api tests into api folder
* replace references to test_assert or test_color with gtest equivalents
* leverage gtest patterns to remove boilerplate code
* dramatically simplify meson.build files and localize variables
* move shared helper code into helpers/ folder
* make test names unique and add gtest protocol for reporting back to
teamcity
Fixes:
* CLDR test keyboards had invalid unicodeset escapes (needs to be raised
upstream also) - was not failing tests because files were not loading
but returning success
Test-bot: skip
This replaces and enables the commented `k_000_null_keyboard` test which
didn't work because keys that are not on any layer don't produce output.
This instead defines a minimal keyboard with just two keys and then tests
typing a key that is on the keyboard followed by a key not on the keyboard.
Test-bot: skip
Production code uses a list of context items, so this change modifies
the tests to also use a list instead of a vector to more closely match
production code.
Addresses code review comment.
Test-bot: skip
This change replaces the reverse iterator loop that holds a stale iterator
across `pop_back()` calls with a pattern that directly accesses
`context.back()` on each iteration. This avoids undefined behavior from
iterator invalidation when mutating the list or vector.
While so far the previous code didn't show problems, it might still access
released memory depending on the implementation. The documentation for
`pop_back()` says "References and iterators to the erased element are
invalidated", so the previous implementation was clearly wrong.
Test-bot: skip
When normalizing, we need to stop processing on an NFC boundary, not an
NFD boundary, to support normalizations such as in Bengali, where
appending `U+09D7` to a context of `U+0995 U+09C7` should result in
`U+0995 U+09CC`.
The specification is unclear on this; see https://unicode-org.atlassian.net/browse/CLDR-19218
This also updates the ldml keyboard unit test suite to support running
in full NFC mode (used in all Engine implementations) as well retaining
the NFD mode (now only used by the debugger).
Side note: the Bengali normalization failure case was picked up by the
improvements to the unit test suite, proving once again that good tests
are so valuable.
Fixes: #15491Fixes: #15505
Follows: #15488
Relates-to: CLDR-19218
Ensure that when a single backspace decomposes the last NFC character in
the app context, the remainder of the 'cluster' is preserved, matching
the implication of the CLDR keyboard specification.
This addresses the behavior in #15487 where the cached context became
out of sync with the app context after deleting an entire NFC cluster
such as ê, which caused a loop ending up with the entire context being
deleted, at which point the loop exited with a fail-safe.
Note that the LDML keyboard tests (ldml.cpp) do not currently exercise
the normalization code; this is a gap that should be addressed to ensure
that we are testing final application behavior.
Fixes: #15487
This substantial refactor reorganizes the header data for sections in
the Core LDML processor. The change was substantial because of
assumptions made about the binary layout of sections. In order to make
the code easier to maintain, safer, and more resilient to future
changes, I opted to make a consistent helper for each section, and copy
header data so that it could be transparently reused.
The principal changes are:
1. Support the version field in the COMP_KMXPLUS_HEADER struct, and
split it into COMP_KMXPLUS_HEADER_17 and COMP_KMXPLUS_HEADER_19
versions.
2. Establish a corresponding COMP_KMXPLUS_XXXX_Helper clas for each
section.
3. Refactor the majority of rawData accesses into using helper
functions, which reduces direct pointer manipulation and adds extra
boundary checks.
A special-case exists for BKSP - it is identical to TRAN, except for
its section ident. In order to avoid a complicated pattern for handling
it, I have special-cased it in one place, adding an overloaded
`get_section_from_sect` function for COMP_KMXPLUS_BKSP.
An opportunity exists to refactor a bit further - reduce direct access
to the binary data (through the COMP_KMXPLUS_XXXX structs, and instead
promote use of the COMP_KMXPLUS_XXXX_Helper classes). This indirection
would reduce duplication of data access and make it cleaner when we
start doing multi-version code.
Next commit will add unit tests for v19 format files.
Build all the .kmx files used for kmn and ldml unit tests in batch
rather than one at a time. The ldml-test-data builds remain unbatched
for now (and are visibly slow!).
Fixes: #13495
- improve an output string
- catch BMP noncharacters besides U+FFFF (H/T @mcdurdin - this was the whole point of the PR) plus test
Fixes: #9446
Co-authored-by: Marc Durdin <marc@durdin.net>
- improve how validation works - a missing section does not mean an invalid section. distingush these.
- propagate errors for invalid sections
- update documentation of required sections
Fixes: #9446
- turn off some asserts- makes untestable
(there are asserts at 'higher levels' such as loading the entire kmx+)
- add a test with a synthesized COMP_KMXPLUS_STRS - a valid and an invalid one
Fixes: #9446
- yes, expand 'other' to all possible combinations
- use ALT and CTRL instead of RALT,LALT and RCTRL,LCTRL in the key list (reduce expansions up to 4x)
Fixes: #12298
Had a real yak shave this morning with disabling assertions in release
builds in our C/C++ code. It turns out that our unit tests use
`assert()` which we intended to use from `test_assert.h`, but in some
cases `cassert` or `assert.h` had been #included after `test_assert.h`,
overriding our special `assert()` macro. The chain of includes is
somewhat hard to puzzle out -- it's often buried several levels deep.
This meant that a release build would drop all test assertions, meaning
most tests passed, unsurprisingly, as there were no assertions left to
fail ... but some tests failed with crashes because we optimized out
important lines such as `assert(some_important_function())`.
I was quite unhappy with this fragility, so I have opted to rename
`assert()` to `test_assert()` in all of our home-grown C/C++ unit tests,
which further highlighted unit tests which were only using the C/C++
`assert()` and not ours, so then had to figure out which unit test
executables needed to have `test_assert` added, and then ... then ...
discovered a bug in `test_color.h`, where we were #including
`io.h`/`unistd.h` inside a `namespace console_color {}` block, which
just happened to be the first ref to those beautiful headers, and thus
(because `#pragma once`) meant that useful little functions like
`access()` were no longer accessible to us in the global namespace.
I have also audited Every Single Call to `assert()` to verify that we do
not do Important Work inside the parentheses, and, apart from those
offending unit tests, now resolved with `test_assert()`, it looks like
all is good.
I would like to present one very well-shaved yak in this commit.
Fixes: #12619
