Colon-pair keywords

[pakal]

Here is the port to GitHub of that previous pull request : https://bitbucket.org/pytest-dev/pytest/pull-request/305/allow-selection-with-myfilepy-myfile/diff#comment-7531336

[pakal]

OK my bad, python2.6 doesn't accept re.sub(..., flags=xxx) argument.
I'll change it to "re.sub("(?i)mr", "", "Mr Bean")" notation asap and update the PR.

[nicoddemus]

Related to #765

[flub]

So continuing the conversation from the original bitbucket PR, does it help if -k understands the node-id? The whole point of the node-id was that it could be copy-pasted, partially copy-pasting it seems like a weird use-case and increases the backwards-compatibility someone who'd like to tackle the python keywords issue with -k would have to worry about.

[RonnyPfannschmidt]

I'm -1 on this one, test ids can be passed without -k

[RonnyPfannschmidt]

Imho We should support exact full name prefix matching and regex matching via the planned keyword expressions

[nicoddemus]

I was +1 on the PR until I found out that you can actually copy/paste ids directly, but now I must say I'm +0 on it as I agree with @flub's points.

[RonnyPfannschmidt]

we should take a look at partial test id's for the following use-case:

-> run all parametrizations of a test functions

[flub]

But as part of -k? I think that should be part of the normal test id
parameter so remain -1 on this
On 22 Jul 2015 19:55, "Ronny Pfannschmidt" [email protected] wrote:

we should take a look at partial test id's for the following use-case:

-> run all parametrizations of a test functions

—
Reply to this email directly or view it on GitHub
#806 (comment).

[RonnyPfannschmidt]

yes, it should be part of the test id selection

i suspect a reorganization of collection<>parametrisation will do some good,
but thats potentially for pytest 3.0

[flub]

So it seems like the consensus is that selection by ID is already suitably covered and should not be done by -k. Thanks for submitting the pull requests anyway!