)]}'
{"README.md":[{"author":{"_account_id":7069,"name":"Joshua Hesketh","email":"josh@nitrotech.org","username":"jhesketh"},"change_message_id":"04246d188cb69c7eb2e4cae6093a6de2c702e9de","unresolved":false,"context_lines":[{"line_number":100,"context_line":"     be considered next to the default bindep.txt."},{"line_number":101,"context_line":"  * `EXTRA_PYDEP` Specify a pydep-* file to add in the container. It would"},{"line_number":102,"context_line":"     be considered next to the default pydep.txt."},{"line_number":103,"context_line":"  * `REGISTRY_SSL` Specify this if you are running an image registry on HTTPS"},{"line_number":104,"context_line":"     holding your requirements image wheels."},{"line_number":105,"context_line":"  * `REGISTRY_SSL_NOVERIFY` Specify this if you are running an image registry"},{"line_number":106,"context_line":"     on HTTPS with self-signed certificates holding your requirements image"}],"source_content_type":"text/x-gfm","patch_set":2,"id":"9fdfeff1_be83b302","line":103,"range":{"start_line":103,"start_character":5,"end_line":103,"end_character":17},"updated":"2019-02-28 11:26:14.000000000","message":"Couldn\u0027t SSL be implied byt the REGISTRY_SSL_NOVERIFY?\nOr otherwise, perhaps this should be \"FORCE_SSL\" or something that avoids falling back to plain http.","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":17068,"name":"Jean-Philippe Evrard","email":"openstack@a.spamming.party","username":"evrardjp"},"change_message_id":"caf400e6f67439f9f34b7c72865244e6fe931418","unresolved":false,"context_lines":[{"line_number":100,"context_line":"     be considered next to the default bindep.txt."},{"line_number":101,"context_line":"  * `EXTRA_PYDEP` Specify a pydep-* file to add in the container. It would"},{"line_number":102,"context_line":"     be considered next to the default pydep.txt."},{"line_number":103,"context_line":"  * `REGISTRY_SSL` Specify this if you are running an image registry on HTTPS"},{"line_number":104,"context_line":"     holding your requirements image wheels."},{"line_number":105,"context_line":"  * `REGISTRY_SSL_NOVERIFY` Specify this if you are running an image registry"},{"line_number":106,"context_line":"     on HTTPS with self-signed certificates holding your requirements image"}],"source_content_type":"text/x-gfm","patch_set":2,"id":"9fdfeff1_48ff6caa","line":103,"range":{"start_line":103,"start_character":5,"end_line":103,"end_character":17},"in_reply_to":"9fdfeff1_be83b302","updated":"2019-02-28 13:27:07.000000000","message":"REGISTRY_FORCE_SSL could be used indeed. Do you think that\u0027s more explicit? Is that a no-go with the current state?","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":7069,"name":"Joshua Hesketh","email":"josh@nitrotech.org","username":"jhesketh"},"change_message_id":"04246d188cb69c7eb2e4cae6093a6de2c702e9de","unresolved":false,"context_lines":[{"line_number":102,"context_line":"     be considered next to the default pydep.txt."},{"line_number":103,"context_line":"  * `REGISTRY_SSL` Specify this if you are running an image registry on HTTPS"},{"line_number":104,"context_line":"     holding your requirements image wheels."},{"line_number":105,"context_line":"  * `REGISTRY_SSL_NOVERIFY` Specify this if you are running an image registry"},{"line_number":106,"context_line":"     on HTTPS with self-signed certificates holding your requirements image"},{"line_number":107,"context_line":"     wheels."},{"line_number":108,"context_line":""}],"source_content_type":"text/x-gfm","patch_set":2,"id":"9fdfeff1_9eaeef7e","line":105,"range":{"start_line":105,"start_character":28,"end_line":105,"end_character":40},"updated":"2019-02-28 11:26:14.000000000","message":"Specify this to what? What if a user does `export REGISTRY_SSL_NOVERIFY\u003dFalse`, how would they expect that to work?\n\nWe should allow these variables to take boolean values. We should also look to avoid the double negative that the above example makes, but I\u0027m fine with the name for now.","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":17068,"name":"Jean-Philippe Evrard","email":"openstack@a.spamming.party","username":"evrardjp"},"change_message_id":"caf400e6f67439f9f34b7c72865244e6fe931418","unresolved":false,"context_lines":[{"line_number":102,"context_line":"     be considered next to the default pydep.txt."},{"line_number":103,"context_line":"  * `REGISTRY_SSL` Specify this if you are running an image registry on HTTPS"},{"line_number":104,"context_line":"     holding your requirements image wheels."},{"line_number":105,"context_line":"  * `REGISTRY_SSL_NOVERIFY` Specify this if you are running an image registry"},{"line_number":106,"context_line":"     on HTTPS with self-signed certificates holding your requirements image"},{"line_number":107,"context_line":"     wheels."},{"line_number":108,"context_line":""}],"source_content_type":"text/x-gfm","patch_set":2,"id":"9fdfeff1_28c0e84f","line":105,"range":{"start_line":105,"start_character":28,"end_line":105,"end_character":40},"in_reply_to":"9fdfeff1_9eaeef7e","updated":"2019-02-28 13:27:07.000000000","message":"I don\u0027t like the double negation. But here I would expect people would not specify it, except to set it to True, which makes it clear for me.\nSo you would prefer a REGISTRY_SSL_VERIFY to True by default, which can get set as \"False\" to skip verification?\n\nFine for that, but I have the impression we are nitpicking here :p","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"}],"scripts/fetch_wheels.py":[{"author":{"_account_id":7069,"name":"Joshua Hesketh","email":"josh@nitrotech.org","username":"jhesketh"},"change_message_id":"04246d188cb69c7eb2e4cae6093a6de2c702e9de","unresolved":false,"context_lines":[{"line_number":63,"context_line":"    return resp.read()"},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"def protocol_detection(registry, protocol\u003d\u0027http\u0027):"},{"line_number":66,"context_line":"    PROTOCOLS \u003d (\u0027http\u0027,\u0027https\u0027)"},{"line_number":67,"context_line":"    index \u003d PROTOCOLS.index(protocol)"},{"line_number":68,"context_line":"    try:"},{"line_number":69,"context_line":"        url \u003d \"{}://{}\".format(protocol, registry)"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_dea1d7aa","line":66,"range":{"start_line":66,"start_character":16,"end_line":66,"end_character":32},"updated":"2019-02-28 11:26:14.000000000","message":"Shouldn\u0027t this (in general) be optimised to try https first as a preference?","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":17068,"name":"Jean-Philippe Evrard","email":"openstack@a.spamming.party","username":"evrardjp"},"change_message_id":"caf400e6f67439f9f34b7c72865244e6fe931418","unresolved":false,"context_lines":[{"line_number":63,"context_line":"    return resp.read()"},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"def protocol_detection(registry, protocol\u003d\u0027http\u0027):"},{"line_number":66,"context_line":"    PROTOCOLS \u003d (\u0027http\u0027,\u0027https\u0027)"},{"line_number":67,"context_line":"    index \u003d PROTOCOLS.index(protocol)"},{"line_number":68,"context_line":"    try:"},{"line_number":69,"context_line":"        url \u003d \"{}://{}\".format(protocol, registry)"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_a89378a2","line":66,"range":{"start_line":66,"start_character":16,"end_line":66,"end_character":32},"in_reply_to":"9fdfeff1_dea1d7aa","updated":"2019-02-28 13:27:07.000000000","message":"I haven\u0027t touched this code, and in fact, I would rather get rid of it in a different patch.","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":7069,"name":"Joshua Hesketh","email":"josh@nitrotech.org","username":"jhesketh"},"change_message_id":"04246d188cb69c7eb2e4cae6093a6de2c702e9de","unresolved":false,"context_lines":[{"line_number":67,"context_line":"    index \u003d PROTOCOLS.index(protocol)"},{"line_number":68,"context_line":"    try:"},{"line_number":69,"context_line":"        url \u003d \"{}://{}\".format(protocol, registry)"},{"line_number":70,"context_line":"        r \u003d urllib2.Request(url)"},{"line_number":71,"context_line":"        resp \u003d urllib2.urlopen(r)"},{"line_number":72,"context_line":"    except (urllib2.URLError,urllib2.HTTPError) as err:"},{"line_number":73,"context_line":"        if err.reason \u003d\u003d \u0027Forbidden\u0027:"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_7e90eb4e","line":70,"range":{"start_line":70,"start_character":20,"end_line":70,"end_character":27},"updated":"2019-02-28 11:26:14.000000000","message":"If REGISTRY_SSL_NOVERIFY was honoured here we could detect the protocol without the changes on line 122","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":17068,"name":"Jean-Philippe Evrard","email":"openstack@a.spamming.party","username":"evrardjp"},"change_message_id":"caf400e6f67439f9f34b7c72865244e6fe931418","unresolved":false,"context_lines":[{"line_number":67,"context_line":"    index \u003d PROTOCOLS.index(protocol)"},{"line_number":68,"context_line":"    try:"},{"line_number":69,"context_line":"        url \u003d \"{}://{}\".format(protocol, registry)"},{"line_number":70,"context_line":"        r \u003d urllib2.Request(url)"},{"line_number":71,"context_line":"        resp \u003d urllib2.urlopen(r)"},{"line_number":72,"context_line":"    except (urllib2.URLError,urllib2.HTTPError) as err:"},{"line_number":73,"context_line":"        if err.reason \u003d\u003d \u0027Forbidden\u0027:"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_c83d9c8e","line":70,"range":{"start_line":70,"start_character":20,"end_line":70,"end_character":27},"in_reply_to":"9fdfeff1_7e90eb4e","updated":"2019-02-28 13:27:07.000000000","message":"not really because we can\u0027t assume if SSL is used or not.\nIf SSL is used (whether or not REGISTRY_SSL_NOVERIFY is used), urllib2 doesn\u0027t throw a URLError nor a HTTPError, and protocol would return the \"protocol\" passed in the function arguments, which by default would be http, except if we change the L122.","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":9963,"name":"Jiri Suchomel","email":"jiri.suchomel@suse.com","username":"jsuchome"},"change_message_id":"e45ca1a0a3b93efe4f7997463d645b7d4a9eb307","unresolved":false,"context_lines":[{"line_number":122,"context_line":"        if os.environ.get(\u0027REGISTRY_SSL\u0027):"},{"line_number":123,"context_line":"            protocol \u003d \u0027https\u0027"},{"line_number":124,"context_line":"        else:"},{"line_number":125,"context_line":"            protocol \u003d protocol_detection(registry)"},{"line_number":126,"context_line":"        kwargs \u003d dict()"},{"line_number":127,"context_line":"        if registry:"},{"line_number":128,"context_line":"            kwargs.update({\u0027registry\u0027: registry})"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_68953e88","line":125,"updated":"2019-02-20 08:20:45.000000000","message":"I think protocol_detection would work correctly, if you\u0027d pass second argument (based on the presence of REGISTRY_SSL).\n\nHowever that method (protocol_detection) looks strange, it would make more sense if it would really try to decect just from the registry URL without hints of what to expect","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":17068,"name":"Jean-Philippe Evrard","email":"openstack@a.spamming.party","username":"evrardjp"},"change_message_id":"06184bfd0d0f72e9b75f3cd8e9e237d45290a7aa","unresolved":false,"context_lines":[{"line_number":122,"context_line":"        if os.environ.get(\u0027REGISTRY_SSL\u0027):"},{"line_number":123,"context_line":"            protocol \u003d \u0027https\u0027"},{"line_number":124,"context_line":"        else:"},{"line_number":125,"context_line":"            protocol \u003d protocol_detection(registry)"},{"line_number":126,"context_line":"        kwargs \u003d dict()"},{"line_number":127,"context_line":"        if registry:"},{"line_number":128,"context_line":"            kwargs.update({\u0027registry\u0027: registry})"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_b9003504","line":125,"in_reply_to":"9fdfeff1_68953e88","updated":"2019-02-21 17:43:31.000000000","message":"the registry is not in form of a url, it\u0027s just an ip/fqdn:port/repo. That is the reason there is detection needed.\n\nBut I think explicit is better than implicit, and in the face of ambiguity, refuse the temptation to guess :D\n\nI think this whole method should disappear anyway.","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":9963,"name":"Jiri Suchomel","email":"jiri.suchomel@suse.com","username":"jsuchome"},"change_message_id":"06ced0c7f4446bdfb6a06fcab59023c29edaa16c","unresolved":false,"context_lines":[{"line_number":122,"context_line":"        if os.environ.get(\u0027REGISTRY_SSL\u0027):"},{"line_number":123,"context_line":"            protocol \u003d \u0027https\u0027"},{"line_number":124,"context_line":"        else:"},{"line_number":125,"context_line":"            protocol \u003d protocol_detection(registry)"},{"line_number":126,"context_line":"        kwargs \u003d dict()"},{"line_number":127,"context_line":"        if registry:"},{"line_number":128,"context_line":"            kwargs.update({\u0027registry\u0027: registry})"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_ed0b571f","line":125,"in_reply_to":"9fdfeff1_b9003504","updated":"2019-02-21 20:23:18.000000000","message":"I agree that protocol_detection looks horrible and should go away. But since we want to keep changes small, wouldn\u0027t it be better to call\n\nprotocol \u003d protocol_detection(registry, \"https\" if os.environ.get(\u0027REGISTRY_SSL\u0027) else \"http\")\n\n(Sorry if my python syntax is wrong)","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":17068,"name":"Jean-Philippe Evrard","email":"openstack@a.spamming.party","username":"evrardjp"},"change_message_id":"7cb7cd10b7beec0068ce478c282a6307e08c4357","unresolved":false,"context_lines":[{"line_number":122,"context_line":"        if os.environ.get(\u0027REGISTRY_SSL\u0027):"},{"line_number":123,"context_line":"            protocol \u003d \u0027https\u0027"},{"line_number":124,"context_line":"        else:"},{"line_number":125,"context_line":"            protocol \u003d protocol_detection(registry)"},{"line_number":126,"context_line":"        kwargs \u003d dict()"},{"line_number":127,"context_line":"        if registry:"},{"line_number":128,"context_line":"            kwargs.update({\u0027registry\u0027: registry})"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_aefb05bc","line":125,"in_reply_to":"9fdfeff1_ed0b571f","updated":"2019-02-21 23:43:51.000000000","message":"I agree it would work, but should we do this? :D\nAnyway, let\u0027s see what other people tell here.","commit_id":"d75b3ff47b3a4a0f5001d3b07297ac5dc2abc31e"},{"author":{"_account_id":7822,"name":"Chris Hoge (not involved in upstream development)","email":"chris@hogepodge.com","username":"hogepodge"},"change_message_id":"5f72997784a8f23697b2282e73e3554a3766ff92","unresolved":false,"context_lines":[{"line_number":25,"context_line":"    print(url)"},{"line_number":26,"context_line":"    try:"},{"line_number":27,"context_line":"        r \u003d urllib2.Request(url\u003durl)"},{"line_number":28,"context_line":"        if os.environ.get(\u0027REGISTRY_INSECURE\u0027, \"False\") \u003d\u003d \"True\":"},{"line_number":29,"context_line":"            resp \u003d urllib2.urlopen(r, context\u003dssl._create_unverified_context())"},{"line_number":30,"context_line":"        else:"},{"line_number":31,"context_line":"            resp \u003d urllib2.urlopen(r)"}],"source_content_type":"text/x-python","patch_set":3,"id":"5fc1f717_4dbd1967","line":28,"updated":"2019-03-05 17:48:57.000000000","message":"Can we make these checks case insensitive, and perhaps even strict  to prevent unexpected default behavior?","commit_id":"27d22da2daf568c82620d09aef6c6845f04cf392"},{"author":{"_account_id":17068,"name":"Jean-Philippe Evrard","email":"openstack@a.spamming.party","username":"evrardjp"},"change_message_id":"3c1920987d2bb874fb7b0b1b05b69db030aff83b","unresolved":false,"context_lines":[{"line_number":25,"context_line":"    print(url)"},{"line_number":26,"context_line":"    try:"},{"line_number":27,"context_line":"        r \u003d urllib2.Request(url\u003durl)"},{"line_number":28,"context_line":"        if os.environ.get(\u0027REGISTRY_INSECURE\u0027, \"False\") \u003d\u003d \"True\":"},{"line_number":29,"context_line":"            resp \u003d urllib2.urlopen(r, context\u003dssl._create_unverified_context())"},{"line_number":30,"context_line":"        else:"},{"line_number":31,"context_line":"            resp \u003d urllib2.urlopen(r)"}],"source_content_type":"text/x-python","patch_set":3,"id":"5fc1f717_a4836656","line":28,"in_reply_to":"5fc1f717_4dbd1967","updated":"2019-03-06 10:27:00.000000000","message":"I don\u0027t really understand, you want to make it strictly equal to \"true\" only, or at the other way around make it more loosely valued?\n\nI don\u0027t mind either. I would use either strict \u003d\u003d \"true\" or use distutils\u0027 strtobool.","commit_id":"27d22da2daf568c82620d09aef6c6845f04cf392"}]}
