search for: xattributes

On Fri, Jan 10, 2014 at 01:33:38PM +0000, Richard W.M. Jones wrote: > The API is now pretty confusing. Each OBool flag is really a > tristate. It can either be "true", "false" or "not specified". > > Therefore I think it should be: > > xattributes:true # copies only xattrs and nothing else > all:true # copies everything > all:true xattributes:false # copies everything except xattrs > > In other words, 'all' changes the default (ie. "not specified") state > of the other flag...

...#39;ve added a "all" argument which would enable every change, overriding > even mode:false. The API is now pretty confusing. Each OBool flag is really a tristate. It can either be "true", "false" or "not specified". Therefore I think it should be: xattributes:true # copies only xattrs and nothing else all:true # copies everything all:true xattributes:false # copies everything except xattrs In other words, 'all' changes the default (ie. "not specified") state of the other flags. To be clearer, the fo...

..., > > > > attached there is a prototype of patch for adding a new > > copy-attributes command. Such command would allow copy the > > attributes of a "file" to> > > another, so for example in guestfish: > > copy-attributes foo bar permissions:true xattributes:false > > > > would only copy the permissions of foo to bar, not copying its > > extended attributes too. > > I think the general idea of the new API is fine. > > More comments about the code below. > > > Just few notes: > > - my first daemon command...

...clude "actions.h" +#include "optgroups.h" GUESTFSD_EXT_CMD(str_file, file); GUESTFSD_EXT_CMD(str_zcat, zcat); @@ -584,3 +585,74 @@ do_filesize (const char *path) return buf.st_size; } + +int +do_copy_attributes (const char *src, const char *dest, int all, int mode, int xattributes, int ownership) +{ + int r; + struct stat srcstat, deststat; + + static const unsigned int file_mask = 07777; + + /* If it was specified to copy everything, manually enable all the flags + * not manually specified to avoid checking for flag || all everytime. + */ + if (all) { + if (!(op...

Hi, attached there is a prototype of patch for adding a new copy-attributes command. Such command would allow copy the attributes of a "file" to another, so for example in guestfish: copy-attributes foo bar permissions:true xattributes:false would only copy the permissions of foo to bar, not copying its extended attributes too. Just few notes: - my first daemon command, so possibly I could be missing something - copy_xattrs is in xattr.c to avoid spreading the usage of xattr API in many places - copy_xattrs does a bit of code...

...3PM +0100, Pino Toscano wrote: > Hi, > > attached there is a prototype of patch for adding a new copy-attributes > command. Such command would allow copy the attributes of a "file" to > another, so for example in guestfish: > copy-attributes foo bar permissions:true xattributes:false > would only copy the permissions of foo to bar, not copying its extended > attributes too. I think the general idea of the new API is fine. More comments about the code below. > Just few notes: > - my first daemon command, so possibly I could be missing something > - copy_...

On Fri, Jan 10, 2014 at 05:36:39PM +0100, Pino Toscano wrote: > +int > +copy_xattrs (const char *src, const char *dest) > +{ > + abort (); > +} Is this called in the case where someone does copy-attributes foo bar xattributes:true on a version of libguestfs with no support for xattrs? It's probably better to silently ignore this case instead of crashing. Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Fedora Windows cross-compiler. Compile Windows programs, test, and build...

On Fri, Jan 10, 2014 at 05:36:39PM +0100, Pino Toscano wrote: [..] This still isn't quite what I meant. My meaning was that mode would be disabled by default (unless all:true). How about: int copy_mode, copy_xattributes, copy_ownership; /* Set defaults. */ if (all) copy_mode = copy_xattributes = copy_ownership = 1; else copy_mode = copy_xattributes = copy_ownership = 0; /* If set in the original struct, copy those settings overriding * the defaults. */ if ((optargs_bitmask & GUESTFS_COPY_ATTRIBUTES_MOD...

...l flag is really a > > tristate. It can either be "true", "false" or "not specified". I see, I did not pay much attention to the use of optargs_bitmask outside the auto-generated RPC stuff the daemon code. > > Therefore I think it should be: > > xattributes:true # copies only xattrs and nothing else > > all:true # copies everything > > all:true xattributes:false # copies everything except xattrs > > > > In other words, 'all' changes the default (ie. "not specified") > &gt...

...QAContact: rsync-qa@samba.org With the git source, the -X flag no-longer works. I suspect this problem was introduced fixing bug #5526. For a simple test case try: sudo mkdir /var2 sudo rsync -aX --link-dest=/var /var/. /var2/. ls -laZ /var2 You should see that /var2 does not have the correct xattributes. -- Configure bugmail: https://bugzilla.samba.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact.

...QAContact: rsync-qa@samba.org With the git source, the -X flag no-longer works. I suspect this problem was introduced fixing bug #5526. For a simple test case try: sudo mkdir /var2 sudo rsync -aX --link-dest=/var /var/. /var2/. ls -laZ /var2 You should see that /var2 does not have the correct xattributes. -- Configure bugmail: https://bugzilla.samba.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the QA contact for the bug, or are watching the QA contact.

...M. Jones wrote: > On Fri, Jan 10, 2014 at 05:36:39PM +0100, Pino Toscano wrote: > > +int > > +copy_xattrs (const char *src, const char *dest) > > +{ > > + abort (); > > +} > > Is this called in the case where someone does > > copy-attributes foo bar xattributes:true > > on a version of libguestfs with no support for xattrs? It's probably > better to silently ignore this case instead of crashing. This version shouldn't be ever called, since it is called when optgroup_linuxxattrs_available returns true, which is not the case in the !HA...

...014 16:53:32 Richard W.M. Jones wrote: > On Fri, Jan 10, 2014 at 05:36:39PM +0100, Pino Toscano wrote: > [..] > > This still isn't quite what I meant. My meaning was that mode > would be disabled by default (unless all:true). OK. > How about: > > int copy_mode, copy_xattributes, copy_ownership; > > /* Set defaults. */ > if (all) > copy_mode = copy_xattributes = copy_ownership = 1; > else > copy_mode = copy_xattributes = copy_ownership = 0; > > /* If set in the original struct, copy those settings overriding > * the defaults. > */ >...

Currently we are unable to properly run guestfsd (the daemon) under valgrind. Attempts to run valgrind inside the appliance have not been successful (see patch 1/7). However we desperately need better valgrind coverage of the daemon, particularly because it is doing a lot of complex parsing of program output. This has been a problem for a long time. A better way to attack this problem is to

...7 +8786,7 @@ This function is used internally when testing the appliance." }; { defaults with name = "copy_attributes"; added = (1, 25, 21); - style = RErr, [Pathname "src"; Pathname "dest"], [OBool "all"; OBool "mode"; OBool "xattributes"; OBool "ownership"]; + style = RErr, [String (Pathname, "src"); String (Pathname, "dest")], [OBool "all"; OBool "mode"; OBool "xattributes"; OBool "ownership"]; shortdesc = "copy the attributes of a path (file/...

Very large but mechanical change to the generator. Rich.

This is a follow-up to the other generator changes in: https://www.redhat.com/archives/libguestfs/2017-February/msg00217.html Rich.

Split up the huge generator/actions.ml into several smaller files. Rich.

The existing mechanism was clunky, slow and used ~ 10 MB of local disk. Rich.