Commits

Show all
Author Commit Message Labels Comments Date
Devin Jeanpierre
Added (untested) support for builtin functions and methods
Devin Jeanpierre
unittest uses MultiParser now -- returns TestSuites instead of TestCases
Devin Jeanpierre
more minor refactoring
Devin Jeanpierre
Refactoring finder
Devin Jeanpierre
Moving towards orthogonalizing loading API.
Devin Jeanpierre
Removed all explicit passing of language; language is part of the test case. It's funny how excellent design seems to spring out of the random short-sighted efforts of improving functionality. I don't ever need to rearchitect, just make the short-sighted decisions that give me just enough flexibility to do what I want. Here the API is significantly cleaner, all because I needed examples to be parsed all at once. I needed examples to be parsed all at once so that they did…
Devin Jeanpierre
All the test functions now use MultiParser
Devin Jeanpierre
MultiParser is no longer broken. DocTests and examples both only have one language per instance now.
Devin Jeanpierre
DocTest sites now only can have one language. If you want more languages, even if the exact same statements are run (polyglot), use multiple DocTests. Also, testmod now uses MultiParser. Also, MultiParser is broken.
Devin Jeanpierre
Moving parser to be compatible with a MultiParser. From now on parsers return lists of DocTests. This is handy in that it allows for there to be multiple DocTests per docstring in the future, as well as just allowing a silly API compatibility between MultiParser and Parser.
Devin Jeanpierre
Small change, going home nowish: DocTestMultiParser can return Doctests, Doctests now have associated languages.
Devin Jeanpierre
Removed dead DocTest attributes: ps1, ps2
Devin Jeanpierre
Added parser that grabs non-overlapping Examples from various languages. parser.DocTestMultiParser is a bit of a hack, but it does the job. This is necessary to be able to handle test case failures distinctly.
Devin Jeanpierre
Parsing no longer dedents text, ever. Where dedenting is desirable, it is performed prior to passing text to the doctest parser. This change is because we want the parser not to munge text it receives -- ideally, we should be able to reconstruct the source text given the parsed text !
Devin Jeanpierre
Changed doctest parsing to use generators rather than lists
Devin Jeanpierre
Quick conversion from list-appending to building for DocTestParser
Devin Jeanpierre
More docs (complete?)
Devin Jeanpierre
Looks like Sphinx autodoc works now -- added new autodocstrings!
Devin Jeanpierre
updated documentation
Devin Jeanpierre
Removed all references to +SH/+CMD; fixed docs up (maybe)
Devin Jeanpierre
Removed simultaneous cmd/sh support
Devin Jeanpierre
Version should reflect expectedstability :|
Devin Jeanpierre
Fixed setup.py omission; doctest2.languages should be installed
Devin Jeanpierre
More documentation on improvements in doctest2
Devin Jeanpierre
Minor comment fix: `output` is matched by traceback parsing too
Devin Jeanpierre
You can now have simultaneous expected-output and expected-exceptions.
Devin Jeanpierre
extra modules for the backwards compatibility test: packages
Devin Jeanpierre
some new compatibility test stuff. run `python3 -m doctest2.tests.cmp stdlib`
Devin Jeanpierre
added custom compile setting
Devin Jeanpierre
Changed output parsing to be done on a per-language basis. This was a very wide-reaching change! Son, I'm impress.
  1. Prev
  2. Next