cordova-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bryan Higgins <br...@bryanhiggins.net>
Subject Re: missing documentation and missing major version bump
Date Wed, 30 Jul 2014 14:55:32 GMT
This came up via an internal client who was using Cordova CLI with their
build script. They were passing in the BB keystore password for release
builds. It was easily resolved by adding --. While unfortunate that this
got through undocumented during a minor update, it's not a huge deal.

I agree with Mark that the separation provides greater flexibility so I
vote for not reverting this. Josh has already updated the help text. We
also need to update the BB10 platform guide for release builds and
deployment to named targets.

Re CI - we do run the unit tests on CI and invested a bunch of time to get
medic running on BB10. This specific test seems to have slipped through the
cracks. I will personally try to be more diligent about keeping that
updated and ensuring release validation gets done. My time has lately been
eaten up by another internal initiative leaving Josh with a lot on his
plate.


On Wed, Jul 30, 2014 at 10:24 AM, Mark Koudritsky <kamrik@google.com> wrote:

> On Wed, Jul 30, 2014 at 1:46 AM, Carlos Santana <csantana23@gmail.com>
> wrote:
>
> > Yep Michal I agree to update cli to pass down extra parameters to
> platform
> > scripts like it used. this allow for greater flexibility in platform
> > scripts and no hardcoded platform concerns in cli. No need to remove nopt
> > if the this can be achieved.
> >
> >
> The list of flags to pass down to run will have to be hard coded in the
> cli, unless we want to go back to the manual position based argument
> parsing.
>
> I don't think we want
> cordova run --verbose
> to mean something different than
> cordova --verbose run
> but I think it's ok to expect that
> cordova run -- --verbose will mean something different.
> And it also allows things like
> cordova --verbose run -- --verbose
>
> The -- separation is exactly what allows the flexibility to pass anything
> to the run and build scripts without telling the cli what to expect or not
> expect after the -- .
>
> But going back to the origins of this discussion. The lack of documentation
> in help.txt about the -- thing is my bad. Thanks to Josh for adding it
> yesterday. But why revert the behavior? (and why only for the pre-4.0
> versions) If webworks is not affected, who is? Are those real people? If
> they have already hit, and possibly solved the problems caused by this
> change, do they want to revert back? Would be glad to hear from them.
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message