~james-page/simplestreams/multihypervisor : changes

~james-page/simplestreams/multihypervisor » Changes from revision 425

From Revision 425 to 406

Rev	Summary	Authors	Date
425	Also set hypervisor_type on stream data so that tools can in... Also set hypervisor_type on stream data so that tools can index by hypervisor_type if need be	James Page	8 years ago
424	Fixup hypervisor_mapping key use Fixup hypervisor_mapping key use	James Page	8 years ago
423	Add configuration option to enable hypervisor mapping in gla... Add configuration option to enable hypervisor mapping in glance mirror	James Page	8 years ago
422	Use constants for disk format mappings Use constants for disk format mappings	James Page	8 years ago
421	Tidy Tidy	James Page	8 years ago
420	Set sane default Set sane default	James Page	8 years ago
419	Fixup disk_format handling Fixup disk_format handling	James Page	8 years ago
418	Fixup lint Fixup lint	James Page	8 years ago
417	Fixup syntax Fixup syntax	James Page	8 years ago
416	Handle disk format and set hypervisor type for glance mirror Handle disk format and set hypervisor type for glance mirror	James Page	8 years ago
415	Only condense tags that are present in all entries. Only condense tags that are present in all entries. There is no way to indicate, at an item level, that a key is not present. This means that in order to be a candidate for compaction, the key must be present in all items. So using set.update() is not correct. We don't want the union of all keys in all items, we want the intersection of those keys.	Scott Moser	8 years ago
414	add json2streams command add json2streams command json2streams accepts a list of JSON files as input, each of those files containing a JSON list of simplestreams "items", as emitted by sstream-query --json. With --juju-format specified, it uses the Juju conventions for writing files (no colons, complete index in index2.json). Otherwise, it uses the standard simplestreams format. As a driveby: * write_streams to not modify its input. * write_streams to accept condense=True argument indicating if any condensing should be done in the data written.	Scott Moser	8 years ago
413	tools/ : xenial will not ship .tar.gz files tools/ : xenial will not ship .tar.gz files The .tar.gz file (compressed sparse tarball of unpartitioned disk image) will not ship in xenial. Modify the is_expected to address that.	Scott Moser	8 years ago
412	tools/toolutil.py: document is_expected much better tools/toolutil.py: document is_expected much better When working on is_expected, I was confused. This change is only: a.) a name change of 'repl' to 'suffix' b.) add description of the function	Scott Moser	8 years ago
411	tools/ : adjust to handle changes in /query data in xenial tools/ : adjust to handle changes in /query data in xenial xenial will not have a -root.tar.gz file, so it wont advertise that file in /query. Adjust the loading of download data to find the right one.	Scott Moser	8 years ago
410	add Xenial Xerus to hard coded list add Xenial Xerus to hard coded list This is not strictly necessary. The fix to get knowledge of Xenial is actually delivered via distro-info-data's SRU under bug 1508759.	Scott Moser	8 years ago
409	Extract generate_simplestreams from tab2streams Extract generate_simplestreams from tab2streams Extract generate_simplestreams from tab2streams, so that it can be reused. Separating out FileNamer may seem strange, but Juju uses a distinct naming scheme for streams: 1. The proper index file is called index2.json (index.json is a cut-down version with only one content-id due to limitations of old jujus) 2. The content files are not named exactly after the content-ids. Instead, ':' is replaced with '-', because ':' is a special character on Windows.	Scott Moser	8 years ago
408	streams data: add 'support_eol' (end of life) streams data: add 'support_eol' (end of life) Each release will now have a 'support_eol' in the data.	Scott Moser	8 years ago
407	sstream-mirror: do not write status info to stdout sstream-mirror: do not write status info to stdout the dry run / report information was going to stdout. That seems wrong.	Scott Moser	8 years ago
406	Add JSON output format to sstream-query. Add JSON output format to sstream-query. adds a new JSON output format to sstream-query, as an alternative to --pretty or the default stringified dicts. Supporting JSON makes sense. It's well-known, expressive, and is also the native storage format of simplestreams. Juju QA are using JSON to generate streams, so --json lets us avoid a translation step.	Scott Moser	8 years ago

Older »