Open denismaier opened 3 years ago
Without any hint of complaint or criticism, could I ask that this and future bugs be represented by a test filed as a pull request to https://github.com/Juris-M/citeproc-js, under fixtures/local
, confirmed to reproduce the bug?
Shouldn't running cslrun -a
reveal the bug?
Anyway... I'm getting strange errors runing the test suite on a newly cloned citeproc-js repo:
Rebundling processor
Error: ENOENT: no such file or directory, scandir 'C:\Users\denis\Desktop\citeproc-js\fixtures\std\processor-tests\humans'
Error: ENOENT: no such file or directory, scandir 'C:\Users\denis\Desktop\citeproc-js\fixtures\std\processor-tests\humans'
at Object.readdirSync (fs.js:891:3)
at checkAll (C:\Users\denis\AppData\Roaming\npm\node_modules\citeproc-test-runner\runtests.js:260:29)
at fetchTestData (C:\Users\denis\AppData\Roaming\npm\node_modules\citeproc-test-runner\runtests.js:310:13)
at bundleValidateTest (C:\Users\denis\AppData\Roaming\npm\node_modules\citeproc-test-runner\runtests.js:660:9)
at C:\Users\denis\AppData\Roaming\npm\node_modules\citeproc-test-runner\runtests.js:862:13
at Object.<anonymous> (C:\Users\denis\AppData\Roaming\npm\node_modules\citeproc-test-runner\runtests.js:875:3)
at Module._compile (internal/modules/cjs/loader.js:1151:30)
at Object.Module._extensions..js (internal/modules/cjs/loader.js:1171:10)
at Module.load (internal/modules/cjs/loader.js:1000:32)
at Function.Module._load (internal/modules/cjs/loader.js:899:14) {
errno: -4058,
syscall: 'scandir',
code: 'ENOENT',
path: 'C:\\Users\\denis\\Desktop\\citeproc-js\\fixtures\\std\\processor-tests\\humans'
}
Any hints?
You will get that error when the repo is cloned without the --recursive
option as shown in the README. To install the submodules and their descendants in an existing repository clone, you can do:
git submodule init;
git submodule update;
Then running cslrun -a
should run and show three errors, from abbreviation failures in legal support that are not relevant to the delimiter issue.
Ok, cslrun works now. (I was getting a bunch of errors when cloning recursively though. Looks like you need to have access to a couple of repos.)
That aside, the test runs as expected. So, if there is a bug, it must be elsewhere...
This is what I'm currently getting with Juris-M and Chicago Author Date (17th edition):
Obviously, there should be a comma.
A user has reported an issue with Chicago Author Date styles. Collapsing of multiple citations is currently incorrect.
According to the test suite, we should get these results:
However, even without locators multiple works by the same author are currently separated with a semicolon rather than with a comma. Earlier that should have worked correctly. Is that a regression?