From 28c97d3104f64ca5b3697130b5c8eee9ee9e2543 Mon Sep 17 00:00:00 2001 From: "https://www.google.com/accounts/o8/id?id=AItOawlmOvQXLmGkQAUOIkPBQtsd9d45NqcX80M" Date: Mon, 3 Mar 2014 20:27:58 +0000 Subject: [PATCH 1/3] --- ...lable_both_here_and_in_another_remote.mdwn | 28 +++++++++++++++++++ 1 file changed, 28 insertions(+) create mode 100644 doc/bugs/whereis_claims_file_is_not_here__44___but_it_is_available_both_here_and_in_another_remote.mdwn diff --git a/doc/bugs/whereis_claims_file_is_not_here__44___but_it_is_available_both_here_and_in_another_remote.mdwn b/doc/bugs/whereis_claims_file_is_not_here__44___but_it_is_available_both_here_and_in_another_remote.mdwn new file mode 100644 index 0000000000..ee69014d27 --- /dev/null +++ b/doc/bugs/whereis_claims_file_is_not_here__44___but_it_is_available_both_here_and_in_another_remote.mdwn @@ -0,0 +1,28 @@ +`git annex whereis` claims that I have zero copies of two particular files. However, it is wrong; both files are present: + +[[!format txt """ +[jkt@vorvan fotky]$ git annex whereis '2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301'* +(merging synced/git-annex into git-annex...) +whereis 2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301.CR2 (0 copies) failed +whereis 2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301.JPG (0 copies) failed +git-annex: whereis: 2 failed +[jkt@vorvan fotky]$ ll 2011-08-13\ Svatba\ Anička\ Fellnerová\ a\ výlet\ s\ Julií/IMG_4301* +lrwxrwxrwx. 1 jkt jkt 331 Aug 13 2011 2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301.CR2 -> ../.git/annex/objects/mg/qq/SHA512E-s19424303--9bdd6917fe85a450b50233a444c1814d407d41b550b66661272728aaf139576cc7c09de158f47a33ff263d7d25253cec193d0e44ed5fb4a1581c28a9a05e75e1.CR2/SHA512E-s19424303--9bdd6917fe85a450b50233a444c1814d407d41b550b66661272728aaf139576cc7c09de158f47a33ff263d7d25253cec193d0e44ed5fb4a1581c28a9a05e75e1.CR2 +lrwxrwxrwx. 1 jkt jkt 329 Aug 13 2011 2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301.JPG -> ../.git/annex/objects/zG/xq/SHA512E-s4744850--da9b1e19f0ecdf08d2f5564232a9c7d4b5ec9fbc9c67033121e5c90468e09e5573ddac42469cfd51fb0d8328dff925830e2d753022365df6d92b115b77831228.JPG/SHA512E-s4744850--da9b1e19f0ecdf08d2f5564232a9c7d4b5ec9fbc9c67033121e5c90468e09e5573ddac42469cfd51fb0d8328dff925830e2d753022365df6d92b115b77831228.JPG +"""]] + +Looking at another repo, the situation is the same: + +[[!format txt """ +jkt@svist /mnt/storage5/fotky[master] $ git annex whereis '2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301'* +whereis 2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301.CR2 (0 copies) failed +whereis 2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301.JPG (0 copies) failed +git-annex: whereis: 2 failed +jkt@svist /mnt/storage5/fotky[master] $ ll '2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301'* +lrwxrwxrwx 1 jkt jkt 331 Mar 3 02:08 2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301.CR2 -> ../.git/annex/objects/mg/qq/SHA512E-s19424303--9bdd6917fe85a450b50233a444c1814d407d41b550b66661272728aaf139576cc7c09de158f47a33ff263d7d25253cec193d0e44ed5fb4a1581c28a9a05e75e1.CR2/SHA512E-s19424303--9bdd6917fe85a450b50233a444c1814d407d41b550b66661272728aaf139576cc7c09de158f47a33ff263d7d25253cec193d0e44ed5fb4a1581c28a9a05e75e1.CR2 +lrwxrwxrwx 1 jkt jkt 329 Mar 3 02:08 2011-08-13 Svatba Anička Fellnerová a výlet s Julií/IMG_4301.JPG -> ../.git/annex/objects/zG/xq/SHA512E-s4744850--da9b1e19f0ecdf08d2f5564232a9c7d4b5ec9fbc9c67033121e5c90468e09e5573ddac42469cfd51fb0d8328dff925830e2d753022365df6d92b115b77831228.JPG/SHA512E-s4744850--da9b1e19f0ecdf08d2f5564232a9c7d4b5ec9fbc9c67033121e5c90468e09e5573ddac42469cfd51fb0d8328dff925830e2d753022365df6d92b115b77831228.JPG +"""]] + +The directory names are valid UTF-8. These are very common on my machine and there is a ton of directories with these funny names here -- all working without any real trouble. + +As far as I know, the file which the links point to is absolutely correct and not corrupted. Looking at the files in the directory chronologically, it also appears that the symlinks point to a correct file. From bf981aa1d03805ffa605784b09ce82717264bc22 Mon Sep 17 00:00:00 2001 From: "https://www.google.com/accounts/o8/id?id=AItOawlmOvQXLmGkQAUOIkPBQtsd9d45NqcX80M" Date: Mon, 3 Mar 2014 20:34:44 +0000 Subject: [PATCH 2/3] --- ...ted_even_though_the_content_has_the_same_hash.mdwn | 11 +++++++++++ 1 file changed, 11 insertions(+) create mode 100644 doc/bugs/variant-_files_are_created_even_though_the_content_has_the_same_hash.mdwn diff --git a/doc/bugs/variant-_files_are_created_even_though_the_content_has_the_same_hash.mdwn b/doc/bugs/variant-_files_are_created_even_though_the_content_has_the_same_hash.mdwn new file mode 100644 index 0000000000..8bf3c3d205 --- /dev/null +++ b/doc/bugs/variant-_files_are_created_even_though_the_content_has_the_same_hash.mdwn @@ -0,0 +1,11 @@ +After (probably, doesn't hepl anymore) adding the same file into multiple clones and syncing them back, I get a couple of these `.variant-` files. The target of these symlinks is different, but the actual file contents of the real data is the same. + +[[!format txt """ +jkt@svist /mnt/storage5/fotky[master] $ ls -al 2011-02-05\ Svatba\ Maruška\ Pavel/jkt/cam1/IMG_8962*variant* +lrwxrwxrwx 1 jkt jkt 335 Mar 3 02:08 2011-02-05 Svatba Maruška Pavel/jkt/cam1/IMG_8962.JPG.variant-847f.cam -> +../../../.git/annex/objects/P4/F7/SHA512E-s5534099--bf29fad838adb755b98351477e9a57c456b66ca393652f3b5df2093fc5448b3eeb0235420a715bfb84fb0cccfbbcea4514d4558467b64a39dc7daaf43c045672.cam/SHA512E-s5534099--bf29fad838adb755b98351477e9a57c456b66ca393652f3b5df2093fc5448b3eeb0235420a715bfb84fb0cccfbbcea4514d4558467b64a39dc7daaf43c045672.cam +lrwxrwxrwx 1 jkt jkt 343 Mar 3 02:08 2011-02-05 Svatba Maruška Pavel/jkt/cam1/IMG_8962.JPG.variant-c8d0.cam -> +../../../.git/annex/objects/38/pG/SHA512E-s5534099--bf29fad838adb755b98351477e9a57c456b66ca393652f3b5df2093fc5448b3eeb0235420a715bfb84fb0cccfbbcea4514d4558467b64a39dc7daaf43c045672.JPG.cam/SHA512E-s5534099--bf29fad838adb755b98351477e9a57c456b66ca393652f3b5df2093fc5448b3eeb0235420a715bfb84fb0cccfbbcea4514d4558467b64a39dc7daaf43c045672.JPG.cam +"""]] + +I don't know what determines the `P4/F7/` vs `38/pG/` directory names, but I would prefer if these duplicates were not created. From d6c69b8d6637a4c0215cb29490fb516162032e53 Mon Sep 17 00:00:00 2001 From: "http://joeyh.name/" Date: Mon, 3 Mar 2014 21:58:21 +0000 Subject: [PATCH 3/3] Added a comment --- ...mment_1_ffc5f79368b8927817e0e35a7a8f057b._comment | 12 ++++++++++++ 1 file changed, 12 insertions(+) create mode 100644 doc/bugs/variant-_files_are_created_even_though_the_content_has_the_same_hash/comment_1_ffc5f79368b8927817e0e35a7a8f057b._comment diff --git a/doc/bugs/variant-_files_are_created_even_though_the_content_has_the_same_hash/comment_1_ffc5f79368b8927817e0e35a7a8f057b._comment b/doc/bugs/variant-_files_are_created_even_though_the_content_has_the_same_hash/comment_1_ffc5f79368b8927817e0e35a7a8f057b._comment new file mode 100644 index 0000000000..3d3531fcec --- /dev/null +++ b/doc/bugs/variant-_files_are_created_even_though_the_content_has_the_same_hash/comment_1_ffc5f79368b8927817e0e35a7a8f057b._comment @@ -0,0 +1,12 @@ +[[!comment format=mdwn + username="http://joeyh.name/" + ip="209.250.56.146" + subject="comment 1" + date="2014-03-03T21:58:20Z" + content=""" +You have 2 different keys, so get variant files. This is expected behavior. + +You're right that the 2 files have the same content. But two different keys were generated for them when they were added in their respective repositories. This can happen when you use the SHA512E (or SHA256E) backend, because it uses the extension as part of the key. One of the files you added had an extension of \".cam\", and the other \".JPG.cam\" + +If there is any bug here, it's perhaps that the \"E\" backends are smart about multi-component file extensions, so will treat eg .tar.gz as an extension, rather than just .gz. If this was not done, both files would have had a .cam extension, and the conflict would not have occurred. But while being less smart would have avoided it in this case, it won't in other cases where files can have different extensions but the same content (for example, \".jpeg\" vs \".jpg\"). +"""]]