https://forge-allura.apache.org/p/allura/tickets/You searched for assigned_to:"masterbunnyfu"enThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7545/The login form return_to param should only accept relative urls, and not external urls. An easy check is that '//' is not in the return_to URL (it matches protocol-less urls too). This will prevent phishing sites from taking advantage the login flow to present a malicious page.Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7545/https://forge-allura.apache.org/p/allura/tickets/7484/UniqueOAuthApplicationName enforces unique app names across all users. But how do you know what other app names people are using? Seems like this should be unique per user, or maybe don't worry about uniqueness at all.Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7484/https://forge-allura.apache.org/p/allura/tickets/7325/*Originally created by:* algorithms Uninitialized repositories should not allow you to for them since you cannot create a merge request for an uninitialized repository.AnonymousThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7325/https://forge-allura.apache.org/p/allura/tickets/7305/Add documentation for SCM classes / dispatch, and do minor cleanups.Cory JohnsThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7305/https://forge-allura.apache.org/p/allura/tickets/7287/The auth handler added in [#6701] needs to be documented.Cory JohnsThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7287/https://forge-allura.apache.org/p/allura/tickets/7274/http://markmail.org/message/rghh5epojmgplb2e Some concerns were raised regarding trademark and copyright notice consistency, and clarity of the various links and pages on the graduation vote thread for Apache Allura.Cory JohnsThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7274/https://forge-allura.apache.org/p/allura/tickets/7229/Would be nice if each section on the profile page could fail gracefully, in addition to fixing root error here. ~~~~ File '/var/local/allura/Allura/allura/ext/user_profile/templates/user_index.html', line 20 in top-level template code {% extends g.theme.master %} File '/var/local/allura/Allura/allura/templates/jinja_master/master.html', line 102 in top-level template code {% block content_base %} File '/var/local/allura/Allura/allura/ext/user_profile/templates/user_index.html', line 49 in block "content_base" {{ section.display() }} File '/var/local/allura/Allura/allura/ext/user_profile/user_main.py', line 268 in display 'auth': AuthenticationProvider.get(request), File '/var/local/allura/ForgeActivity/forgeactivity/main.py', line 268 in prepare_context activity.obj.project = activity_obj.project if activity_obj else None AttributeError: 'Commit' object has no attribute 'project' ~~~~Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7229/https://forge-allura.apache.org/p/allura/tickets/7215/Since https://builds.apache.org/job/Allura/552/ the test suite has been timing out. It is also failing on some SourceForge-internal CI builds. It even times out after 2.5 hrs.Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7215/https://forge-allura.apache.org/p/allura/tickets/7204/Changes split out from [#7097]: * Sort projects by name, show mod date * I think we should also still show social network accounts on the profile page, by default. SourceForge could possibly disable them if so desired (and its entry forms too, I suppose?)Tim Van SteenburghThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7204/https://forge-allura.apache.org/p/allura/tickets/7181/On every page, jinja_master/nav_menu.html calls `c.project.admins()` When that calls `users_with_named_role` it shouldn't have to query all of the project roles since it only cares about the one with the specific name.Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7181/https://forge-allura.apache.org/p/allura/tickets/7178/Sometimes when using ./run_tests with multiple processes, I get lots of these errors: ~~~~ File "/var/local/env-allura/lib/python2.7/site-packages/nose-1.3.0-py2.7.egg/nose/case.py", line 381, in setUp try_run(self.inst, ('setup', 'setUp')) File "/var/local/env-allura/lib/python2.7/site-packages/nose-1.3.0-py2.7.egg/nose/util.py", line 469, in try_run return func() File "/home/dbrondsema/dbrondsema-1019/forge/AlluraTest/alluratest/controller.py", line 129, in setUp setup_functional_test(app_name=self.application_under_test)) File "/home/dbrondsema/dbrondsema-1019/forge/AlluraTest/alluratest/controller.py", line 83, in setup_functional_test setup_basic_test(config, app_name) File "/home/dbrondsema/dbrondsema-1019/forge/AlluraTest/alluratest/controller.py", line 75, in setup_basic_test while M.MonQTask.run_ready('setup'): File "/home/dbrondsema/dbrondsema-1019/forge/Allura/allura/model/monq_model.py", line 239, in run_ready task() File "/home/dbrondsema/dbrondsema-1019/forge/Allura/allura/model/monq_model.py", line 267, in __call__ self.result = func(*self.args, **self.kwargs) File "/home/dbrondsema/dbrondsema-1019/forge/Allura/allura/tasks/index_tasks.py", line 66, in add_artifacts solr.add(solr_updates) AttributeError: 'NoneType' object has no attribute 'add' ~~~~Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7178/https://forge-allura.apache.org/p/allura/tickets/7134/Users should be able to fork https://forge-allura.apache.org/p/allura/git/ Currently you can make a fork, but there's no way to get your code or to push it back. The 'allura' project on our server there is basically a readonly copy with pointers to https://git-wip-us.apache.org/repos/asf?p=incubator-allura.git For forks (and any other type of local project on the server) to work, we could run git services on the allura-vm host. And probably need some code/config changes to support having a different checkout URL for the 'allura' repo vs other repos.Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7134/https://forge-allura.apache.org/p/allura/tickets/7130/`Blob.context()` (which is called when on the file view page) calls `Blob.next_commit` but the result is not used. `Blob.next_commit` walks down the commit tree to find the next commit that modified the file, which for rarely modified files can result in a lot of useless work. `Blob.prev_commit` is more efficient, but is also not used. Remove both, as well as associated tests in `Allura/allura/tests/unit/test_repo.py`, and ensure there is no visible impact to the UI of the file view page.Cory JohnsThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7130/https://forge-allura.apache.org/p/allura/tickets/7128/The "Browse Commits" interface can be incredibly slow, especially for larger repositories. Override the `commit_browser_data` method in ForgeSVN to get data directly from svn instead of doing lots of mongo lookups and data manipulation. That implementation in Allura core seems much more complex than is necessary for SVN anyway, since SVN is always a single linear list of commits (no "real" branches, merges, or multiple heads to consider)Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7128/https://forge-allura.apache.org/p/allura/tickets/7124/https://sourceforge.net/nf/admin/task_manager/view/52ebe3f21be1ce2af7b78b31 Validate Trac URLs so we don't get 404s or spurious NoneType errors during import.Cory JohnsThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7124/https://forge-allura.apache.org/p/allura/tickets/7119/Especially for tests, template overrides are not always desired. Add a config switch to disable overrides globally, and enable it in `test.ini`.Cory JohnsThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7119/https://forge-allura.apache.org/p/allura/tickets/7118/Like [#6895] for Google Code, our github importer should check to see if a user/repo pair is valid before importing (and erroring on a 404 later)Dave BrondsemaThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7118/https://forge-allura.apache.org/p/allura/tickets/7103/The stale project flushing issue that was fixed for repositories in [#7005] is still an issue for other long-running tool imports. Any time an artifact is saved, the `Artifact.__metaclass__.before_save` method updates `c.project.last_updated` which causes `c.project` to become dirty and, when flushed, overwrite any other changes to the project from another source. For project imports, this can cause the metadata import (or even manual updates to the project metadata) to be overwritten by the tool imports.Cory JohnsThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7103/https://forge-allura.apache.org/p/allura/tickets/7097/Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7097/https://forge-allura.apache.org/p/allura/tickets/7085/~~~~ File '/var/local/allura/ForgeActivity/forgeactivity/main.py', line 120 in feed 'title': 'Activity for %s' % data['followee'].shortname, AttributeError: 'User' object has no attribute 'shortname' ~~~~Dave BrondsemaThu, 20 Aug 2015 22:07:00 -0000https://forge-allura.apache.org/p/allura/tickets/7085/https://forge-allura.apache.org/p/allura/tickets/7063/Cory JohnsThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7063/https://forge-allura.apache.org/p/allura/tickets/7047/On Allura's own ticket tracker when I do a bulk edit, it lately has been not updating solr with the ticket changes. This has happened a few times, but I haven't been able to reproduce on a test project or locally. Here's the tasks that run in such a case: https://sourceforge.net/p/allura/pastebin/52d057cbea406b5dc88d20ec And here's the tasks that run when it does work properly: https://sourceforge.net/p/allura/pastebin/52d057d404161f105486b39c In the case of it not working, the sendmail events and logging of ticket version commits show that the tickets were found. I can even find the corresponding zarkov events that ArtifactSessionExtension.after_flush sends out, so all I can figure is that "arefs" was falsey. But i don't see how that can happen.Dave BrondsemaThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7047/https://forge-allura.apache.org/p/allura/tickets/7026/`def follow` in `forgeactivity/main.py` should require POST. And templates and tests should be changed to send posts (and don't forget the csrf token).Dave BrondsemaThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7026/https://forge-allura.apache.org/p/allura/tickets/7007/[forge:site-support:#6309] >Following up on the cygwin-ports upgrade: >I have been unable to create Git repos via the REST API whose mount_point param contain an underscore, even though such repo names are legal and can be created via the admin/tools web interface. The cause appears to be allura/ext/admin/admin_main.py (line 776 in git master): > if not h.re_tool_mount_point.match(mount_point) or c.project.app_instance(mount_point) is not None: return {'success': False, 'info': 'Incorrect mount point name, or mount point already exists.' } >Should that not be something like this instead: > if c.project.app_instance(mount_point) is not None: return {'success': False, 'info': 'Mount point already exists.' } if not h.re_relaxed_tool_mount_point.match(mount_point): return {'success': False, 'info': 'Incorrect mount point name.' } >IOW, separate the errors for easier diagnosis, and more importantly, use re_relaxed_tool_mount_point to match the web interface. Chris TsaiThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7007/https://forge-allura.apache.org/p/allura/tickets/7005/It appears that the allura.tasks.repo_tasks.clone task flushes the Project record at the end, and since that task can take a very long time, any updates (e.g. to project description) get lost when it saves the old project document back to mongo. I haven't duplicated this in development, but here's what happened on sourceforge: * Google Code import was used; `import_project_info` task ran at 2013/12/17 04:54:38 https://sourceforge.net/nf/admin/task_manager/view/52afd90dc4314373398546ad * Repo clone started at 2013/12/17 04:54:39 https://sourceforge.net/nf/admin/task_manager/view/52afd90e34309d0fdc251462 * short description was manually set, at 2013-12-17 06:07:16 (per Audit Log) * Repo clone finished at 2013/12/17 07:32:09 and currently the short description is empty. I didn't find any other tasks with the specific project_id that completed after 06:07:16, so I believe it's the repo clone task at fault.Dave BrondsemaThu, 20 Aug 2015 22:07:10 -0000https://forge-allura.apache.org/p/allura/tickets/7005/