diff options
13 files changed, 160 insertions, 0 deletions
diff --git a/doc/bugs/__34__Configuring_Jabber_Account__34___fails_with_a___34__Network_unreachable__34___error./comment_1_6d821af99ab3c83a5b0f52d3713ab8e2._comment b/doc/bugs/__34__Configuring_Jabber_Account__34___fails_with_a___34__Network_unreachable__34___error./comment_1_6d821af99ab3c83a5b0f52d3713ab8e2._comment new file mode 100644 index 000000000..c2dd31f17 --- /dev/null +++ b/doc/bugs/__34__Configuring_Jabber_Account__34___fails_with_a___34__Network_unreachable__34___error./comment_1_6d821af99ab3c83a5b0f52d3713ab8e2._comment @@ -0,0 +1,10 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 1" + date="2013-09-25T18:27:24Z" + content=""" +Sounds like the SRV lookup is failing. Does `git-annex version` list either DNS or ADNS in the build flags? + +Does `host -t SRV _xmpp-client._tcp.gmail.com` work? +"""]] diff --git a/doc/bugs/__96__git_annex_import__96___does_not_work_on_other_git_annex_repositories/comment_1_94ccd548c084286163eeb2af1ddc18e3._comment b/doc/bugs/__96__git_annex_import__96___does_not_work_on_other_git_annex_repositories/comment_1_94ccd548c084286163eeb2af1ddc18e3._comment new file mode 100644 index 000000000..4de9cc10d --- /dev/null +++ b/doc/bugs/__96__git_annex_import__96___does_not_work_on_other_git_annex_repositories/comment_1_94ccd548c084286163eeb2af1ddc18e3._comment @@ -0,0 +1,8 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 1" + date="2013-09-25T18:21:25Z" + content=""" +Import skips symlinks and other non-regular files. It would work if the source repository was in direct mode. +"""]] diff --git a/doc/bugs/git_annex_indirect_can_fail_catastrophically/comment_1_0b085e7e8c8e364f479574bc00c7c394._comment b/doc/bugs/git_annex_indirect_can_fail_catastrophically/comment_1_0b085e7e8c8e364f479574bc00c7c394._comment new file mode 100644 index 000000000..68814881d --- /dev/null +++ b/doc/bugs/git_annex_indirect_can_fail_catastrophically/comment_1_0b085e7e8c8e364f479574bc00c7c394._comment @@ -0,0 +1,21 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 1" + date="2013-09-25T18:57:25Z" + content=""" +Worse than being stuck partway converted, it fails in such a way that the file you can't write to is left stuck in .git/annex/objects/ without a symlink pointint to it. + +Here is how to recover: + +1. run `git annex direct` +2. run `git annex indirect` +3. run `git annex direct` +4. run `git annex indirect` +5. run `git revert HEAD` +6. run `git annex direct` +7. fix the permission of the file +8. run `git annex indirect` + +Please don't ask me why this works, but it will.. +"""]] diff --git a/doc/bugs/git_version_in_prebuilt_linux_tarball_is_outdated/comment_1_2a5a07498df9d38531d4570f7b463b9a._comment b/doc/bugs/git_version_in_prebuilt_linux_tarball_is_outdated/comment_1_2a5a07498df9d38531d4570f7b463b9a._comment new file mode 100644 index 000000000..5287ea935 --- /dev/null +++ b/doc/bugs/git_version_in_prebuilt_linux_tarball_is_outdated/comment_1_2a5a07498df9d38531d4570f7b463b9a._comment @@ -0,0 +1,8 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 1" + date="2013-09-25T18:18:49Z" + content=""" +The tarballs are built on Debian stable in order to have an old enough libc to work most places. So I am limited to what is available in stable and backports. Once there is a backport available of git, I will use it. +"""]] diff --git a/doc/bugs/merge_causes_out_of_memory_on_large_repos/comment_5_e8998716107e7ae8d0e8d332812517ad._comment b/doc/bugs/merge_causes_out_of_memory_on_large_repos/comment_5_e8998716107e7ae8d0e8d332812517ad._comment new file mode 100644 index 000000000..272dc2fd2 --- /dev/null +++ b/doc/bugs/merge_causes_out_of_memory_on_large_repos/comment_5_e8998716107e7ae8d0e8d332812517ad._comment @@ -0,0 +1,14 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 5" + date="2013-09-25T18:36:39Z" + content=""" +That doesn't look very big, I merge one 3x that large on a 128 mb machine. + +I think you will need to either email me privately so I can get a copy of your repository to investigate with ... or you can try to investigate on your own. + +I think the first things I would try to debug this are to look over `git annex merge --debug` and see if I see anything unusual, and then I would probably `git checkout git-annex` in the repository, and wc -l on all the files and see if any file has a lot of lines, or is otherwise very large. + +If that found nothing, my next step would be to rebuild git-annex from source with memory profiling enabled, as explained in this book, and try to get a memory profiling graph that explained what was using up the memory. <http://book.realworldhaskell.org/read/profiling-and-optimization.html> +"""]] diff --git a/doc/forum/Import_options.txt b/doc/forum/Import_options.txt new file mode 100644 index 000000000..543d1a4ec --- /dev/null +++ b/doc/forum/Import_options.txt @@ -0,0 +1,14 @@ +Thank you for adding import options to handle duplicates. Very handy when consolidating data from various sources. + +Can deletion of the source files be decoupled from annex duplication/deduplication options? For example, I would like to import source files without deleting them and at the same time do not import duplicates. + +Better yet, since deletion of source files is potentially dangerous, a delete option could be required for deletion to be performed. Example: + +git annex import --deduplicate --delete_all_source_files +git annex import --deduplicate --delete_source_duplicates + +Also, it would be great to have import "status" option which goes over files to be imported and logs their status ( to be imported, duplicate etc. ) without actually performing any changes. It would be great for testing and trial runs. + +I hope the above make sense. It would make import feature more flexible. + +Cheers, diff --git a/doc/forum/can_I_only_add_my_own_files__63__/comment_1_767d647af9d0345f337338d6319071fa._comment b/doc/forum/can_I_only_add_my_own_files__63__/comment_1_767d647af9d0345f337338d6319071fa._comment new file mode 100644 index 000000000..80efaf04b --- /dev/null +++ b/doc/forum/can_I_only_add_my_own_files__63__/comment_1_767d647af9d0345f337338d6319071fa._comment @@ -0,0 +1,10 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 1" + date="2013-09-25T18:42:00Z" + content=""" +git-annex needs to be able to lock down files to ensure that nobody can write to them, and to do this it needs to remove the write bit, and you can't remove the write bit from a file you don't own. + +Note that if you configure git's core.sharedRepository when making a repository (git init --shared), then all files in both git and git-annex will be group writable. Put you and the other person you wanted to be able to write to the file in a group, and you can both access the repository. So that's the right way to do it. +"""]] diff --git a/doc/forum/git-annex___38___ikiwiki_experiment/comment_3_fbbd47c3dbe8de24b0df664e4afd5cb8._comment b/doc/forum/git-annex___38___ikiwiki_experiment/comment_3_fbbd47c3dbe8de24b0df664e4afd5cb8._comment new file mode 100644 index 000000000..63919bd22 --- /dev/null +++ b/doc/forum/git-annex___38___ikiwiki_experiment/comment_3_fbbd47c3dbe8de24b0df664e4afd5cb8._comment @@ -0,0 +1,14 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 3" + date="2013-09-25T17:03:29Z" + content=""" +Interesting experiment. + +I don't know why you don't want to let git-annex sync its data to $gitdir. + +The symlinks could be occuring because of a bug in direct mode. (I have fixed many past bugs that caused that.) But just as likely it's because ikiwiki will run `git pull` in the srcdir. + +I think it would make more sense to use the underlay plugin and keep your annexed repository in a separate underlay. This would guarantee ikiwiki doesn't run git commands in there, and would ensure that nothing done with the web interface could mess with the annex. +"""]] diff --git a/doc/forum/git-annex___38___ikiwiki_experiment/comment_4_55da5c3c41c13b08590ce1ff8117cef6._comment b/doc/forum/git-annex___38___ikiwiki_experiment/comment_4_55da5c3c41c13b08590ce1ff8117cef6._comment new file mode 100644 index 000000000..6299326a8 --- /dev/null +++ b/doc/forum/git-annex___38___ikiwiki_experiment/comment_4_55da5c3c41c13b08590ce1ff8117cef6._comment @@ -0,0 +1,23 @@ +[[!comment format=mdwn + username="https://www.google.com/accounts/o8/id?id=AItOawk3HGoDpnOPob5jOjvIootmkve1-nCpRiI" + nickname="Kalle" + subject="comment 4" + date="2013-09-25T18:28:40Z" + content=""" +@Joey + + > I don't know why you don't want to let git-annex sync its data to $gitdir. + +Well neither do I! :) It seemed to be the way to avoid duplicating data while still having the images picked up by the ikiwiki album plugin. Wouldn't the files in the $gitdir end up duplicated in $srcdir? + + > The symlinks could be occuring because of a bug in direct mode. + > (I have fixed many past bugs that caused that.) But just as likely + > it's because ikiwiki will run git pull in the srcdir. + +When you mention it I've had similar problems with my vfat usb annex repos. Using the post-receive merge hook to make files visible for non git-annex devices about town. Nothing I can reliably recreate but I will keep my eye out for bugs. + + > I think it would make more sense to use the underlay plugin and keep + > your annexed repository in a separate underlay. + +Yep that would be ideal. For my usecase the album plugin is vital and I can't understand how album would pick up and deal with images in an underlay dir. This is a bit OT for this site though most of my questions are ikiwiki related. +"""]] diff --git a/doc/special_remotes/xmpp/comment_6_8f0b5bba1271d031a67e7f0c175d67d5._comment b/doc/special_remotes/xmpp/comment_6_8f0b5bba1271d031a67e7f0c175d67d5._comment new file mode 100644 index 000000000..750e0874a --- /dev/null +++ b/doc/special_remotes/xmpp/comment_6_8f0b5bba1271d031a67e7f0c175d67d5._comment @@ -0,0 +1,8 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 6" + date="2013-09-25T18:13:24Z" + content=""" +If you're not getting an \"XMPPClient: connected\", then my guess would be that your git-annex build's XMPP is screwed up somehow. For example, if it hung forever when connecting to the XMPP server, it would never get as far as printing that message. (If it tried and failed to connect, you'd get a message about the connection having failed.) +"""]] diff --git a/doc/todo/redundancy_stats_in_status/comment_2_686ced0684d10511caf07953c64cd5b6._comment b/doc/todo/redundancy_stats_in_status/comment_2_686ced0684d10511caf07953c64cd5b6._comment new file mode 100644 index 000000000..a4711b2a3 --- /dev/null +++ b/doc/todo/redundancy_stats_in_status/comment_2_686ced0684d10511caf07953c64cd5b6._comment @@ -0,0 +1,10 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 2" + date="2013-09-25T18:03:55Z" + content=""" +`git annex status .` or otherwise running it with a directory has recently started walking all the location logs for all files in the directory in order to display variance from configured numcopies. It would be easy to add a redundancy counter to that. + +It would slow down the global status when not passed a directory to add redundancy info there. Maybe local is enough? +"""]] diff --git a/doc/todo/wishlist:_display_name_of_object_when_addWatcher_gets_a_permission_denied/comment_1_d2665e7347689b520d37561cfddf0aa8._comment b/doc/todo/wishlist:_display_name_of_object_when_addWatcher_gets_a_permission_denied/comment_1_d2665e7347689b520d37561cfddf0aa8._comment new file mode 100644 index 000000000..de0528855 --- /dev/null +++ b/doc/todo/wishlist:_display_name_of_object_when_addWatcher_gets_a_permission_denied/comment_1_d2665e7347689b520d37561cfddf0aa8._comment @@ -0,0 +1,8 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 1" + date="2013-09-25T18:47:13Z" + content=""" +This is an exception from the inotify library, which is what contains the `addWatch` function. I catch and display the exception. Since `addWatch` is only passed a directory to watch, the most I could do is tack on the name of the directory when displaying the exception. That does not seem likely to be much help? +"""]] diff --git a/doc/todo/wishlist:_use_hardlinks_for_local_clones/comment_1_85064fafe472a5bd395d60ce8f7acb56._comment b/doc/todo/wishlist:_use_hardlinks_for_local_clones/comment_1_85064fafe472a5bd395d60ce8f7acb56._comment new file mode 100644 index 000000000..4ef5f8414 --- /dev/null +++ b/doc/todo/wishlist:_use_hardlinks_for_local_clones/comment_1_85064fafe472a5bd395d60ce8f7acb56._comment @@ -0,0 +1,12 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="4.152.108.220" + subject="comment 1" + date="2013-09-25T17:14:28Z" + content=""" +git-annex uses cp --reflink=auto. So on a filesystem supporting COW file copies, like btrfs, `git annex get` will not use any disk space when getting from the same filesystem. + +I do not like the idea of using hardlinks, because changing the file in one repository would change it in the other, which may not be desired. + +[[union_mounting]] seems to cover this item pretty well, so I will close this as a duplicate. +"""]] |