~smoser/simplestreams/add-aliases : changes

~smoser/simplestreams/add-aliases » Changes from revision 424

From Revision 424 to 405

Rev	Summary	Authors	Date
424	add 'aliases' data as a comma separated list to products. add 'aliases' data as a comma separated list to products. Bug: https://bugs.launchpad.net/cpc-core/+bug/1550421 This branch introduces a new "aliases" field as a comma separated list of alias in a string json entry. Its content is as follow: - release codename - the first letter of the release codename - the version number - the aliases "lts" and "default" if the current active LTS - "devel" if the current development release	Scott Moser	8 years ago
423	Enable better support for multi-hypervisor regions in OpenSt... Enable better support for multi-hypervisor regions in OpenStack 1) Determine correct disk format. 2) Set hypervisor_type on images based on ftype field.	Scott Moser	8 years ago
422	dump_data: dump json data with sorted keys dump_data: dump json data with sorted keys make dump_data use sorted keys for dictionaries. The desire is to reduce differences in output that would otherwise be the same.	Scott Moser	8 years ago
421	tools/make-test-data: Fix missing lxc/lxd files for ppc64el ... tools/make-test-data: Fix missing lxc/lxd files for ppc64el builds. We had a whitelist on ppc64el ever since the first ppc64el builds. Removing that whitelist means we will now get correct file information for all ppc64el builds.	Scott Moser	8 years ago
420	Fix TestUrlContentSourcea to test with requests, add build-d... Fix TestUrlContentSourcea to test with requests, add build-depends Fix the test case to use Requests, then fix the build failure that actually testing it caused. Then, support skipping the requests tests if it is not present.	Scott Moser	8 years ago
419	sanitize size parameter to read on Readers sanitize size parameter to read on Readers This fixes the read() calls in Urllib2UrlReader, RequestsUrlReader, and the new FileReader so that they all support: read(-1) read(None) read() and pass the correct value down to the python below it. Essentially, python3 prefers read(size=None) while python2 prefers read(size-1) if the size is passed explicitly rather than using read().	Scott Moser	8 years ago
418	fix lintian warnings, and apply pull back distro package bui... fix lintian warnings, and apply pull back distro package build changes debian/control: - bump standards version - add build-depends on python-mock, python-yamnl, python3-mock - add build-depends on dh-python per sbuild output recommendation debian/copyright: fix license string agplv3 to AGPL-3 debian/rules: set SS_REQUIRE_DISTRO_INFO = 0 during build	Scott Moser	8 years ago
417	fix a bug where --no-verify doesn't work with signed streams... fix a bug where --no-verify doesn't work with signed streams. Although the assignment of cmd was done only if checked was True, cmd was used when checked was not True. This produced a stack trace with UnboundLocalError. This branch fixes it and introduces tests based on test_signed_data.	Scott Moser	8 years ago
416	tools: xenial has .tar.gz files for the moment. tools: xenial has .tar.gz files for the moment. .tar.gz are expected to go away in xenial, but are present for the moment. As other things (maas image build process) needs them for now, we should index them.	Scott Moser	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
405	sstream-mirror: fix trace if .data is not present. sstream-mirror: fix trace if .data is not present. when reading from the mirror, don't bother reading the non '.data' path. As returning that as a product indicates that it has all been done.	Scott Moser	8 years ago

Older »