#6226 NORM Never A: xo-get activity fails to start on update.1 build 690

[Zarro Boogs per Child]

#6226: xo-get activity fails to start on update.1 build 690
------------------------+---------------------------------------------------
  Reporter:  DanKrejsa  |       Owner:  jg                 
      Type:  defect     |      Status:  new                
  Priority:  normal     |   Milestone:  Never Assigned     
 Component:  distro     |     Version:                     
Resolution:             |    Keywords:  rainbow-integration
  Verified:  0          |    Blocking:                     
 Blockedby:             |  
------------------------+---------------------------------------------------
Changes (by mstone):

 * cc: mstone (added)
  * keywords:  => rainbow-integration


Comment:

 First - the log looks strange because' it's colored. Read it with 'less
 -R' and it will be very pretty.

 Second - the log indicates that Rainbow's
 [http://dev.laptop.org/git?p=users/mstone/security;a=blob;f=rainbow/rainbow/inject.py;hb=83eea6528df6a65d9fed508344019cc1e14b24bd#l207
 check_bundle_id()] assertion is failing. This means that you're passing
 Rainbow a bundle-name containing characters it doesn't expect. In this
 case, the '-' is to blame.

 This restriction arose because originally, the bundle_ids represented both
 D-Bus interfaces and D-Bus object paths at the same time. According to the
 [http://dbus.freedesktop.org/doc/dbus-specification.html#message-protocol-
 names D-Bus spec], '-' characters are not allowed in interface names or
 method names.

 As it stands now, I'm implementing a slightly more restrictive check which
 requires that bundle_ids consist of period-separated blocks of
 alphanumerics. I could be convinced that this 'accepted language' isn't
 right, either because it's too small or because it's not precisely
 identical with the one specified by the D-Bus spec, but frankly, I'm not
 sure it's worth it.

 Would renaming the 'xo-get' activity to 'XOGet' (or similar) be
 acceptable?

-- 
Ticket URL: <http://dev.laptop.org/ticket/6226#comment:1>
One Laptop Per Child <http://dev.laptop.org>
OLPC bug tracking system