#1729 - Testing

[trungtruong1]

Link to the issue: #1729

[IshaanVerma1130]

Passed all tests. Yay.

[ZuxingGit]

https://pygments.org/docs/lexerdevelopment/#how-to-test-your-lexer this is link the repo owner gave us. Everyone should read it.

[trungtruong1]

Guys, I updated the build.yml file to run CI for every new commit push.

The original build.yml requires you to push to the master branch to run CI (or run for pull requests)

[adluu2k1]

It seems that our fix also affects some C/Cpp based language (see pygments/lexers/c_like.py) and we will have to update the outputs of these lexers as part of the PR. But before adding them to the PR, we need to check whether these outputs are correct or not.
An ideal process that I would suggest is: --update-goldens everything in tests folder -> use git to identify changed output files -> if the output file is not from a c/cpp based lexer, undo the change and consider it a failed test case. If it is -> check these output files whether they are lexed correctly. If not, undo the change and consider it a failed test case.

[ZuxingGit]

Hello, everyone. I found a problem that our previous fixing change something from Keyword/Type/Name into Name.Function,
which is not correct. Shown below:

tests/examplefiles/cpp/functions.cpp
tests/examplefiles/cpp/functions.cpp.output , these files are for checking this problem

And more:

[ZuxingGit]

Here, we shouldn't combine AnsiGenerator:: together

Same changes happend in: tests/examplefiles/mql/example.mqh.output

[ZuxingGit]

Should it keep the same as before?

[ZuxingGit]

getNextProgramCharDistance in example.cpp. Should it be a Name.Function?

[ZuxingGit]

tests/examplefiles/nesc/IPDispatchP.nc

[ZuxingGit]

Here, one ":" should be seen as Punctuation rather than Operator

[ZuxingGit]

tests/examplefiles/promela/sample.pml, who knows promela's syntax

[ZuxingGit]

tests/examplefiles/swig/swig_java.swg.output is full of these problems like above mentioned:

[IshaanVerma1130]

It seems that our fix also affects some C/Cpp based language (see pygments/lexers/c_like.py) and we will have to update the outputs of these lexers as part of the PR. But before adding them to the PR, we need to check whether these outputs are correct or not. An ideal process that I would suggest is: --update-goldens everything in tests folder -> use git to identify changed output files -> if the output file is not from a c/cpp based lexer, undo the change and consider it a failed test case. If it is -> check these output files whether they are lexed correctly. If not, undo the change and consider it a failed test case.

The author has explained us a method to run test for only the lexer we have modified. I believe we should be doing that.

[trungtruong1]

But if we do that, it will fail CI. For example, our C++ lexer will create an output for tests/examplefiles/swig/swig_java.swg. This output will be compared with tests/examplefiles/swig/swig_java.swg.output and will return fail. tests/examplefiles/swig/swig_java.swg is using the C++ lexer.