Menu
▾
▴
Re: [ntfs-3g-devel] ntfs-3g I/O errors on "unsupported reparse point"
|
From: Jean-Pierre A. <jea...@wa...> - 2020-03-03 14:05:54
|
Pascal wrote: > hi Jean-Pierre, > > if it is not too complicated to implement, what changes should be made > to the source code of ntfs-3g so that these pseudo files systematically > appear as broken links rather than regular files ? Well, the complexity lies in the fact that the needed information is not in the directory which is being listed. You have to open the files referenced in the directory and delegate the identification of their type to the plugin which knows about it. I have uploaded to https://jp-andre.pagesperso-orange.fr/ntfs-3g_ntfsprogs-2017.3.23AA.0.tgz the alpha version of an upgraded ntfs-3g doing that, for you to test. You may also want to test a plugin variant which try to locate the executable designated by the execlink. This obviously implies dropping the arguments, so the plugin is as useless as the previous one. See : https://jp-andre.pagesperso-orange.fr/execlink110.zip Note : this plugin can only be used with the proposed alpha version of ntfs-3g. Jean-Pierre > > regards, Pascal. > > Le ven. 28 févr. 2020 à 12:08, Pascal > <pat...@gm... > <mailto:pat...@gm...>> a écrit : > > > (the object is returned as a normal file because there is no way > to say it has an unknown type) > > I suppose that in both cases (ls and find) the informations are > returned by ntfs-3g : is that right ? > from what I understand, these pseudo files are more like links > (broken links to the cloud) than regular files : if you agree with > this point of view, wouldn't it be more relevant to announce them as > such ? > > Pascal > > Le jeu. 27 févr. 2020 à 16:40, Jean-Pierre André > <jea...@wa... > <mailto:jea...@wa...>> a écrit : > > Pascal wrote: > > I have a little question about these special files... > > > > the ls command report MicrosoftEdge and python as (broken) > links : > > > > # ls -l /test/AppData/Local/Microsoft/WindowsApps/ > > total 0 > > drwxrwxrwx 1 root root 0 27 févr. 07:34 Backup > > drwxrwxrwx 1 root root 0 27 févr. 07:39 > > Microsoft.DesktopAppInstaller_8wekyb3d8bbwe > > drwxrwxrwx 1 root root 0 27 févr. 07:34 > > Microsoft.MicrosoftEdge_8wekyb3d8bbwe > > *l*rwxrwxrwx 2 root root 132 27 févr. 07:34 MicrosoftEdge.exe -> > > > 'C:\Windows\system32\SystemUWPLauncher.exe(Microsoft.MicrosoftEdge_8wekyb3d8bbwe,Microsoft.MicrosoftEdge_8wekyb3d8bbwe!MicrosoftEdge)' > > *l*rwxrwxrwx 1 root root 228 27 févr. 07:39 python.exe -> > 'C:\Program > > > Files\WindowsApps\Microsoft.DesktopAppInstaller_1.0.30251.0_x64__8wekyb3d8bbwe\AppInstallerPythonRedirector.exe(Microsoft.DesktopAppInstaller_8wekyb3d8bbwe,Microsoft.DesktopAppInstaller_8wekyb3d8bbwe!PythonRedirector)' > > *l*rwxrwxrwx 2 root root 228 27 févr. 07:39 python3.7.exe -> > 'C:\Program > > > Files\WindowsApps\Microsoft.DesktopAppInstaller_1.0.30251.0_x64__8wekyb3d8bbwe\AppInstallerPythonRedirector.exe(Microsoft.DesktopAppInstaller_8wekyb3d8bbwe,Microsoft.DesktopAppInstaller_8wekyb3d8bbwe!PythonRedirector)' > > *l*rwxrwxrwx 1 root root 228 27 févr. 07:39 python3.exe -> > 'C:\Program > > > Files\WindowsApps\Microsoft.DesktopAppInstaller_1.0.30251.0_x64__8wekyb3d8bbwe\AppInstallerPythonRedirector.exe(Microsoft.DesktopAppInstaller_8wekyb3d8bbwe,Microsoft.DesktopAppInstaller_8wekyb3d8bbwe!PythonRedirector)' > > > > but find command report them as regular files : > > > > # find /test/AppData/Local/Microsoft/WindowsApps/ -type l > > /(nothing is returned)/ > > # find /test/AppData/Local/Microsoft/WindowsApps/ -maxdepth 1 > -type f > > /test/AppData/Local/Microsoft/WindowsApps/MicrosoftEdge.exe > > /test/AppData/Local/Microsoft/WindowsApps/python.exe > > /test/AppData/Local/Microsoft/WindowsApps/python3.7.exe > > /test/AppData/Local/Microsoft/WindowsApps/python3.exe > > > This is because "ls" and "find" do not rely on the same data. > "ls" relies on the object attributes returned by stat(), and > "find" relies on the data stored in directories returned by > readdir(). > > In this situation you are accessing a normal directory which > contains unexpected data (the object is returned as a normal > file because there is no way to say it has an unknown type). > > This can be fixed by analyzing deeper, but I am not keen on > doing so until some useful meaning (which could be different > from a symlink) can be assigned in Linux to such object. > > Jean-Pierre > > > > > which normally should not be the case : > > > > # ls -l /tmp/test/ > > total 0 > > *l*rwxrwxrwx 1 root root 1 27 févr. 10:36 broken.link -> x > > -rw-r--r-- 1 root root 0 27 févr. 10:35 file > > *l*rwxrwxrwx 1 root root 4 27 févr. 10:36 link -> file > > # find /tmp/test/ -type f > > /tmp/test/file > > # find /tmp/test/ -type l > > /tmp/test/broken.link > > /tmp/test/link > > > > Le jeu. 27 févr. 2020 à 09:56, Pascal > > <pat...@gm... > <mailto:pat...@gm...> > > <mailto:pat...@gm... > <mailto:pat...@gm...>>> a > écrit : > > > > hi Jean-Pierre, > > > > the plugin also works perfectly with Archlinux x64. > > > > thanks again, regards, Pascal. > > > > Le mer. 26 févr. 2020 à 18:08, Pascal > <pat...@gm... > <mailto:pat...@gm...> > > <mailto:pat...@pu... > <mailto:pat...@gm...>>> a > écrit : > > > > hi Jean-Pierre, > > > > the plugin works perfectly with Pop_OS (Debian Ubuntu > like) x64 > > : error I/Os no longer occur and files simply appear > as broken > > links. > > I'll be testing shortly with Archlinux... > > > > regards, Pascal. > > > > Le mar. 25 févr. 2020 à 13:58, Pascal > > <pat...@pu... > <mailto:pat...@gm...> > > > <mailto:pat...@pu... > <mailto:pat...@gm...>>> a > > écrit : > > > > I haven't tested the suggested plugin yet, but, > considering > > the results displayed, I'll be sure to look into > it :-) > > > > thanks for everything and especially for ntfs-3g > > > > Le mar. 25 févr. 2020 à 13:46, Jean-Pierre André > > > <jea...@pu... > <mailto:jea...@pu...> > > > <mailto:jea...@pu... <mailto:jea...@pu...>>> > > a écrit : > > > > Pascal wrote: > > > this is more of a "cosmetic" problem : the > > input/output errors displayed > > > are confusing and may not be displayed. > > > > > > # ll > > > /mnt/sdb2/Users/Test/AppData/Local/Microsoft/WindowsApps/ > > > ls: impossible d'accéder à > > > > > > '/mnt/sdb2/Users/Test/AppData/Local/Microsoft/WindowsApps/MicrosoftEdge.exe': > > > > > Erreur d'entrée/sortie > > > ls: impossible d'accéder à > > > > > > '/mnt/sdb2/Users/Test/AppData/Local/Microsoft/WindowsApps/python.exe': > > > > > Erreur d'entrée/sortie > > > ls: impossible d'accéder à > > > > > > '/mnt/sdb2/Users/Test/AppData/Local/Microsoft/WindowsApps/python3.7.exe': > > Erreur > > > d'entrée/sortie > > > ls: impossible d'accéder à > > > > > > '/mnt/sdb2/Users/Test/AppData/Local/Microsoft/WindowsApps/python3.exe': > > > > > Erreur d'entrée/sortie > > > total 0 > > > drwxrwxrwx 1 root root 0 10 sept. 14:00 Backup > > > drwxrwxrwx 1 root root 0 10 sept. 14:03 > > > Microsoft.DesktopAppInstaller_8wekyb3d8bbwe > > > drwxrwxrwx 1 root root 0 10 sept. 14:00 > > > Microsoft.MicrosoftEdge_8wekyb3d8bbwe > > > -????????? ? ? ? ? ? MicrosoftEdge.exe > > > -????????? ? ? ? ? ? python.exe > > > -????????? ? ? ? ? ? python3.7.exe > > > -????????? ? ? ? ? ? python3.exe > > > > > > are the unsupported reparse point dates > displayed by > > the ll command > > > correct when the missing module is created > symbolically ? > > > > Did you install the plugin I suggested ? > > > > This is what you should get (well > approximately, please > > ignore > > the reformatting by the mailer), and > obviously the links > > lead > > to no usable files (as ntfs-3g cannot start a > remote Windows > > executable). > > > > If you do not get this, please post the > output of the > > script from : > > https://jp-andre.pagesperso-orange.fr/junctions.html#other > > > > > > $ ls -l > > > /win10/Users/Jean-Pierre/AppData/Local/Microsoft/WindowsApps > > total 0 > > drwx------ 1 linux linux 0 Jun 14 2019 Backup > > lrwxrwxrwx 2 linux linux 198 Jan 4 13:37 > > GameBarElevatedFT_Alias.exe -> > > 'C:\Program > > > Files\WindowsApps\Microsoft.XboxGamingOverlay_3.36.6003.0_x64__8wekyb3d8bbwe\GameBarElevatedFT.exe(Microsoft.XboxGamingOverlay_8wekyb3d8bbwe,Microsoft.XboxGamingOverlay_8wekyb3d8bbwe!App)' > > drwx------ 1 linux linux 0 Jan 4 13:35 > > Microsoft.DesktopAppInstaller_8wekyb3d8bbwe > > lrwxrwxrwx 2 linux root 132 Nov 1 11:11 > > MicrosoftEdge.exe -> > > > 'C:\WINDOWS\system32\SystemUWPLauncher.exe(Microsoft.MicrosoftEdge_8wekyb3d8bbwe,Microsoft.MicrosoftEdge_8wekyb3d8bbwe!MicrosoftEdge)' > > drwx------ 1 linux linux 0 Nov 1 11:11 > > Microsoft.MicrosoftEdge_8wekyb3d8bbwe > > drwx------ 1 linux linux 0 Jan 4 13:37 > > Microsoft.XboxGamingOverlay_8wekyb3d8bbwe > > lrwxrwxrwx 1 linux linux 228 Jan 4 13:35 > python3.exe -> > > 'C:\Program > > > Files\WindowsApps\Microsoft.DesktopAppInstaller_1.0.32912.0_x64__8wekyb3d8bbwe\AppInstallerPythonRedirector.exe(Microsoft.DesktopAppInstaller_8wekyb3d8bbwe,Microsoft.DesktopAppInstaller_8wekyb3d8bbwe!PythonRedirector)' > > lrwxrwxrwx 1 linux linux 228 Jan 4 13:35 > python.exe -> > > 'C:\Program > > > Files\WindowsApps\Microsoft.DesktopAppInstaller_1.0.32912.0_x64__8wekyb3d8bbwe\AppInstallerPythonRedirector.exe(Microsoft.DesktopAppInstaller_8wekyb3d8bbwe,Microsoft.DesktopAppInstaller_8wekyb3d8bbwe!PythonRedirector)' > > > > > > > > > > # ll > > > /mnt/sdb2/Users/Test/AppData/Local/Microsoft/WindowsApps/ > > > total 0 > > > drwxrwxrwx 1 root root 0 10 sept. 14:00 Backup > > > drwxrwxrwx 1 root root 0 10 sept. 14:03 > > > Microsoft.DesktopAppInstaller_8wekyb3d8bbwe > > > drwxrwxrwx 1 root root 0 10 sept. 14:00 > > > Microsoft.MicrosoftEdge_8wekyb3d8bbwe > > > lrwxrwxrwx 2 root root 25 10 sept. 14:00 > > MicrosoftEdge.exe -> > > > 'unsupported reparse point' > > > lrwxrwxrwx 1 root root 25 10 sept. 14:03 > python.exe > > -> 'unsupported > > > reparse point' > > > lrwxrwxrwx 2 root root 25 10 sept. 14:03 > > python3.7.exe -> 'unsupported > > > reparse point' > > > lrwxrwxrwx 1 root root 25 10 sept. 14:03 > python3.exe > > -> 'unsupported > > > reparse point' > > > > > > Pascal > > > > > > Le mar. 25 févr. 2020 à 10:55, Jean-Pierre > André > > > > > > <jea...@pu... <mailto:jea...@pu...> > > > <mailto:jea...@pu... <mailto:jea...@pu...>> > > > > > > > > <mailto:jea...@pu... <mailto:jea...@pu...> > > > <mailto:jea...@pu... <mailto:jea...@pu...>>>> > > a écrit : > > > > > > Pascal wrote: > > > > hi everyone, > > > > > > > > as suggested by the package > maintainer, I'm > > transferring to you > > > an open > > > > ticket on Archlinux bug tracker : > > > https://bugs.archlinux.org/task/63737 > > > > > > > > regards. > > > > > > The reparse point this ticket is > referring to is > > a link for > > > executing some cloud Windows > application (this is > > at least how > > > I understand it) and this leads to an > > "unsupported reparse point" > > > error, because I do not know what kind of > > meaningful support I > > > could provide. > > > > > > Why would you access this reparse > point in the > > first place ? > > > > > > Anyway I have uploaded a minimal plugin to > > > > https://jp-andre.pagesperso-orange.fr/execlink.zip > > > This will show the parameters > available in the > > reparse data, > > > and might give some information for > somebody to > > describe > > > what useful support might be expected. > > > > > > Jean-Pierre > > > > |
