We are actively working to verify that magnum-ui works with the adjusted requirements.txt, and as soon as we have confirmed this change is non-disruptive, I will be ready to approve the FFE.
Adrian > On Feb 7, 2017, at 4:54 PM, Richard Jones <r1chardj0...@gmail.com> wrote: > > It looks like Magnum-UI only has one xstatic package in their > requirements that isn't already in Horizon's requirements (and > therefore is superfluous), and that's xstatic-magic-search, which has > been replaced in Horizon by pulling magic search into the Horizon tree > (we forked because maintaining our own extensions against the package > was getting out of hand - we'd basically rewritten a large proportion > of the code). > > I would recommend that the Magnum-UI project remove all xstatic > packages from requirements.txt > > > Richard > > On 7 February 2017 at 14:17, Tony Breeds <t...@bakeyournoodle.com> wrote: >> On Tue, Feb 07, 2017 at 10:39:41AM +1100, Richard Jones wrote: >>> Hi requirements team, >>> >>> We've had a downstream packager come to us with issues packaging the >>> Horizon RC as described in this bug report: >>> >>> https://bugs.launchpad.net/horizon/+bug/1662180 >>> >>> The issues stems from the requirements file having several xstatic >>> package minimum versions specified that are no longer compatible with >>> Horizon, and the RDO build system honors those minimum version >>> specifications, and boom! >> >> This is a specific case of OpenStack provides poor tools for >> testing/validating >> minimum requirements. This is a thing we started trying to fix in Ocata but >> the work is slow going :( I'm a little confused how this wasn't caught >> sooner >> by RDO (given they would appear to have been testing the minimums for >> xstatic-*) >> >>> Rob Cresswell has proposed a patch to bump those minimum versions up >>> to the versions specified in upper-constraints.txt: >>> >>> https://review.openstack.org/#/c/429753 >> >> That review seems to adjust all Xstatic packages where the minimu != the >> constrained version which is probably more than is required but it doesn't >> actually increase the knock-on effects so it seems like a good idea to me :) >> >> Looking at the projects that are affected by Rob's review: >> >> Package : xstatic-angular [xstatic-angular>=1.3.7] (used by 3 projects) >> Package : xstatic-angular-bootstrap >> [xstatic-angular-bootstrap>=0.11.0.2] (used by 3 projects) >> Package : xstatic-angular-gettext [xstatic-angular-gettext>=2.1.0.2] >> (used by 3 projects) >> Package : xstatic-bootstrap-scss [xstatic-bootstrap-scss>=3.1.1.1] >> (used by 3 projects) >> Package : xstatic-d3 [xstatic-d3>=3.1.6.2] (used by 3 projects) >> Package : xstatic-font-awesome [xstatic-font-awesome>=4.3.0] (used by 3 >> projects) >> Package : xstatic-jasmine [xstatic-jasmine>=2.1.2.0] (used by 3 >> projects) >> Package : xstatic-jsencrypt [xstatic-jsencrypt>=2.0.0.2] (used by 3 >> projects) >> Package : xstatic-rickshaw [xstatic-rickshaw>=1.5.0] (used by 3 >> projects) >> Package : xstatic-smart-table [xstatic-smart-table!=1.4.13.0,>=1.4.5.3] >> (used by 3 projects) >> Package : xstatic-term-js [xstatic-term-js>=0.0.4.1] (used by 3 >> projects) >> openstack/horizon [tc:approved-release] >> openstack/karbor-dashboard [] >> openstack/magnum-ui [] >> >> >> Package : xstatic-bootswatch [xstatic-bootswatch>=3.3.5.3] (used by 1 >> projects) >> openstack/horizon [tc:approved-release] >> >> And obviously RDO >> >> This will mean that Horizon will need an RC2, and any packaging/distro >> testing >> for horizon (and plugins/dashboards) will need to be restarted (iff said >> testing was done with an xstatic package not listed in >> upper-constraaints.txt[1]) >> >> I tried to determine the impact on magnum-ui and karbor-dashboard and AFAICT >> they're already using constraints. The next thing to look at is the release >> model which is: >> magnum-ui: >> type: horizon-plugin >> model: cycle-with-intermediary >> karbor-dashboard: >> type: unknown >> model: unknown >> >> I think this means it's safe grant this FFE as the affected plugins aren't >> necessarily in a stabilisation phase. >> >> So as far as I can see we have 2 options: >> 1. Do nothing: there will be other cases that minimums are not functional. >> RDO have tools and data to fix this in there own repos so we're not >> actually blocking them >> 2. Take the patch, and accept the knock on effects. >> >> I'm okay with taking this FFE if Karbor and Magnum PTLs sign off here (or on >> the review) >> >>> Additionally to the above I will be proposing a patch to Horizon's >>> documented processes to ensure that when an xstatic upper-constraints >>> version is bumped we also bump the minimum version in >>> global-requirements to avoid this sort of thing in the future. >> >> Cool. That'll help >> >> Yours Tony. >> >> [1] We've communicated that u-c should be the source here before >> >> __________________________________________________________________________ >> OpenStack Development Mailing List (not for usage questions) >> Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe >> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev >> > > __________________________________________________________________________ > OpenStack Development Mailing List (not for usage questions) > Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe > http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
