Merge lp:~jcsackett/launchpad/package-pickers-navigation-jumps into lp:launchpad

[j.c.sackett]

[r=benji][bug=698020,698024] Fixes a few minor UI issues with package picker.

Summary
=======
Package pickers have a variety of visual issues.

* Navigation links change position as you navigate through batches because the picker can resize for longer package names
* BinaryAndSourcePackageNames aren't recognized as a type of Package name, so they get the bullet point sprite css, instead of package sprite css.

This fixes those two issues.

Proposed Fix
============
* Set the picker widget to use a fixed width, so that it doesn't resize as the batch loads.
* Update the code selecting the sprite css to provide the package sprite for IBinaryAndSourcePackageName

Preimplementation
=================
Spoke with Curtis Hovey about the CSS fixes for the picker.

Implementation
==============
lib/lp/app/browser/tales.py
---------------------------
* Updated the code that provides sprite_css to provide the appropriate sprite for IBInaryAndSourcePackageNames

lib/canonical/launchpad/icing/style-3-0.css
-------------------------------------------
* Changed yui3-picker's width to 40%, to closely mimic its previous 'auto' behavior but have a fixed width at a given window size.

Tests
=====
No new tests, as these were trivial cosmetic changes.

QA
==
Follow through the instructions on the linked bugs, which provide "What should happen" examples for QA.

Lint
====
make lint was not generating output for me. I will get it running and fix any lint before landing.

[Benji York (benji)]

The building of JavaScript via a page template in form-picker-macros.pt
looks like a JS injection hazard.

In particular the strings inserted verbatim from the view. The JS
should really be just a string literal (or even better, moved to its own
file) and the data can be exposed to the browser using the JSON request
cache (see IJSONRequestCache).

[j.c.sackett (jcsackett)]

I concur, but I think that's a separate bug; I proposed this based on a rollback of something else since I needed that rollback to be able to do this, but as it has since landed I think the diff is out of date and confusing the issue.

I'm going to update and resubmit this MP.

[Benji York (benji)]

review: Approve (code)

=== modified file 'lib/canonical/launchpad/icing/style-3-0.css'
--- lib/canonical/launchpad/icing/style-3-0.css	2011-06-09 13:49:54 +0000
+++ lib/canonical/launchpad/icing/style-3-0.css	2011-06-20 15:52:47 +0000
@@ -1939,6 +1939,10 @@
     }
 
 /* Picker styles */
+.yui3-picker {
+    width: 40%;
+    }
+
 .yui3-picker-results div.badge {
     position: absolute;
     top: 3px;

=== modified file 'lib/lp/app/browser/tales.py'
--- lib/lp/app/browser/tales.py	2011-06-01 13:25:21 +0000
+++ lib/lp/app/browser/tales.py	2011-06-20 15:52:47 +0000
@@ -64,6 +64,9 @@
 from lp.soyuz.enums import ArchivePurpose
 from lp.soyuz.interfaces.archive import IPPA
 from lp.soyuz.interfaces.archivesubscriber import IArchiveSubscriberSet
+from lp.soyuz.interfaces.binarypackagename import (
+    IBinaryAndSourcePackageName,
+    )
 from lp.registry.interfaces.distribution import IDistribution
 from lp.registry.interfaces.distributionsourcepackage import (
     IDistributionSourcePackage,
@@ -741,6 +744,8 @@
             return 'sprite branch'
         elif ISpecification.providedBy(context):
             return 'sprite blueprint'
+        elif IBinaryAndSourcePackageName.providedBy(context):
+            return 'sprite package-source'
         return None
 
     def default_logo_resource(self, context):