[TIP] Disable keyboard shortcuts in html coverage reports?

Ben Finney ben+python at benfinney.id.au
Tue Mar 4 18:28:35 PST 2014


Ned Batchelder <ned at nedbatchelder.com> writes:

> The right thing to do would be to contact the authors and get them to
> change their stance toward their own software. Once they did, then we
> could stop vendoring it.

Yes, that's what I've been advocating. Sorry if I communicated that
poorly.

But along with that, we do need to deal with the current situation in
the meantime.

I'm very loath to defy Debian policy on bundled third-party code,
especially having worked with you (Ned) to address it correctly. So I'll
resist anything that regresses our stance on that.

A possible strategy:

* Ned, as a developer directly dependent on ‘jquery.hotkeys.js’, you
  have IMO the strongest clout with the developers of that library.
  Please communicate with them to understand how breaking backward
  compatibility is undesirable without a clear migration path, and to
  perhaps address this in a future release (and be more considerate of
  dependent developers in future).

* Thomi, your assistance would be appreciated in figuring out what
  change in ‘jquery.hotkeys.js’ has broken some assumption in Coverage.

* If the assumption can be mended to work with both version “0.8” and
  version “0.8+” of that library, good. Coverage can be released in a
  version that works with that library before and after the change.

* If not, and some larger change is required to work with version “0.8+”
  of the library, then we can declare Coverage has a versioned
  dependency on the library for particular versions of Coverage. Some
  future release of the Debian package can thereby refuse to install
  unless a compatible version of the library is available.

That's an explicit spelling-out of what is really rather routine issue
in management of library dependency.

It's common wisdom that releasing a new version of a library needs to
take care with existing code that uses previous versions of the library;
it's a shame that so many JS developers have yet to learn about this,
and appear oblivious of the cost of expecting people to be as cavalier
about new releases as they are.

-- 
 \      “Yesterday I saw a subliminal advertising executive for just a |
  `\                                           second.” —Steven Wright |
_o__)                                                                  |
Ben Finney




More information about the testing-in-python mailing list