)]}'
{"zuul/driver/github/githubconnection.py":[{"author":{"_account_id":4146,"name":"Clark Boylan","email":"cboylan@sapwetik.org","username":"cboylan"},"change_message_id":"c5d4f6444b8e327771ddcd51e45c55d834148fe3","unresolved":false,"context_lines":[{"line_number":1160,"context_line":""},{"line_number":1161,"context_line":"        keys \u003d set()"},{"line_number":1162,"context_line":"        pattern \u003d \u0027 OR \u0027.join(change.uris)"},{"line_number":1163,"context_line":"        query \u003d \u0027%s type:pr is:open in:body\u0027 % pattern"},{"line_number":1164,"context_line":"        # Repeat the search for each installation id (project)"},{"line_number":1165,"context_line":"        for installation_project in installation_projects:"},{"line_number":1166,"context_line":"            github \u003d self.getGithubClient(installation_project)"}],"source_content_type":"text/x-python","patch_set":2,"id":"3fa7e38b_4d443873","line":1163,"updated":"2019-11-26 00:39:55.000000000","message":"Does this mean that we treat any cross referenced PR as a depends on even if there is no Depends-On flag? Seems that we only search for the url of this change in other changes. I think that may contribute to the issue you are seeing in that it is too greedy in its match.\n\nMaybe we can relax the check below if we add an explicit check for depends-on?","commit_id":"6f35b2cc436255396cb1f944c23d31b078859873"},{"author":{"_account_id":16068,"name":"Tobias Henkel","email":"tobias.henkel@bmw.de","username":"tobias.henkel"},"change_message_id":"f06d6471963274a5970dda9e73b20b62afdf15a0","unresolved":false,"context_lines":[{"line_number":1160,"context_line":""},{"line_number":1161,"context_line":"        keys \u003d set()"},{"line_number":1162,"context_line":"        pattern \u003d \u0027 OR \u0027.join(change.uris)"},{"line_number":1163,"context_line":"        query \u003d \u0027%s type:pr is:open in:body\u0027 % pattern"},{"line_number":1164,"context_line":"        # Repeat the search for each installation id (project)"},{"line_number":1165,"context_line":"        for installation_project in installation_projects:"},{"line_number":1166,"context_line":"            github \u003d self.getGithubClient(installation_project)"}],"source_content_type":"text/x-python","patch_set":2,"id":"3fa7e38b_fc167b58","line":1163,"in_reply_to":"3fa7e38b_4d443873","updated":"2019-12-04 19:25:52.000000000","message":"Yes, looks like, but while maybe reducing the error cases refining the search won\u0027t solve the underlying problem. So I think we could refine it but that would be a separate change.","commit_id":"6f35b2cc436255396cb1f944c23d31b078859873"},{"author":{"_account_id":4146,"name":"Clark Boylan","email":"cboylan@sapwetik.org","username":"cboylan"},"change_message_id":"c5d4f6444b8e327771ddcd51e45c55d834148fe3","unresolved":false,"context_lines":[{"line_number":1174,"context_line":"                sha \u003d pr.get(\u0027head\u0027).get(\u0027sha\u0027)"},{"line_number":1175,"context_line":"                key \u003d (proj, num, sha)"},{"line_number":1176,"context_line":""},{"line_number":1177,"context_line":"                trusted, project \u003d tenant.getProject(proj)"},{"line_number":1178,"context_line":"                # ignore projects zuul doesn\u0027t know about"},{"line_number":1179,"context_line":"                if not project:"},{"line_number":1180,"context_line":"                    continue"}],"source_content_type":"text/x-python","patch_set":2,"id":"3fa7e38b_8d4eb093","line":1177,"range":{"start_line":1177,"start_character":53,"end_line":1177,"end_character":57},"updated":"2019-11-26 00:39:55.000000000","message":"This value should be subpath/project which getProject supports fine. However, should we be using the fully qualified canonical name here including the github url as well? That way if a project with the same name is somehow in the same tenant from two different sources we avoid conflicts?\n\nTo do that I think you can replace proj here with url less the scheme? It could also be that I am overthinking this.","commit_id":"6f35b2cc436255396cb1f944c23d31b078859873"},{"author":{"_account_id":16068,"name":"Tobias Henkel","email":"tobias.henkel@bmw.de","username":"tobias.henkel"},"change_message_id":"f06d6471963274a5970dda9e73b20b62afdf15a0","unresolved":false,"context_lines":[{"line_number":1174,"context_line":"                sha \u003d pr.get(\u0027head\u0027).get(\u0027sha\u0027)"},{"line_number":1175,"context_line":"                key \u003d (proj, num, sha)"},{"line_number":1176,"context_line":""},{"line_number":1177,"context_line":"                trusted, project \u003d tenant.getProject(proj)"},{"line_number":1178,"context_line":"                # ignore projects zuul doesn\u0027t know about"},{"line_number":1179,"context_line":"                if not project:"},{"line_number":1180,"context_line":"                    continue"}],"source_content_type":"text/x-python","patch_set":2,"id":"3fa7e38b_17e9f017","line":1177,"in_reply_to":"","updated":"2019-12-04 19:25:52.000000000","message":"good catch, we should use the canonical name here, but not from the url but from self.canonical_hostname. You can construct a Project object here using proj and self.source and let the Project object calculate the canonical name.","commit_id":"6f35b2cc436255396cb1f944c23d31b078859873"}]}