VIM and NTFS streams
Alex
I understand that these bugs and feature requests are "nice to have",
but anyway...
VIM is not working very well with NTFS streams. When you edit file,
say, "test.txt:aaa",
it complains about backup during save (E510). If you edit file
"test.txt:my.txt", VIM opens
and edits file test_txt (but with correct stream ).
Also, it would be very nice if NTFS streams could be shown during
command line
autocompletion. Say, you have file test.txt with streams aaa and bbb.
On command line
you write ":e test.txt:", and after pressing tab you see test.txt:aaa
and test.txt:bbb, like
normal file names.
Best regards,
Alex
Alexei Alexandrov
2 notes here:
1. I think NTFS streams is useless feature. I've never seen any
practical example of the usefulness of this.
2. I'm not sure there is any API to enumerate streams inside a given file.
--
Alexei Alexandrov
Craig Barkhouse
> 2 notes here:
>
> 1. I think NTFS streams is useless feature. I've never seen any
> practical example of the usefulness of this.
For my own purposes, I agree. However, some users and some applications will find uses for streams (same can be said of any feature). It would be much more useful if more filesystems supported it.
> 2. I'm not sure there is any API to enumerate streams inside a given file.
FindFirstStreamW()
FindNextStreamW()
Bram Moolenaar
Alex Jakushev wrote:
Try setting the 'backupcopy' option to "yes".
I thought Vim did copy the streams, but I suppose this doesn't work when
you edit one specific stream. If you want to look at it: function
copy_infostreams() in src/os_win32.c
NTFS streams are mostly restricted to MS-Windows applications, and
rarely used. I don't think it's a good idea to support them in Vim
directly.
--
BEDEVERE: And what do you burn, apart from witches?
FOURTH VILLAGER: ... Wood?
BEDEVERE: So why do witches burn?
SECOND VILLAGER: (pianissimo) ... Because they're made of wood...?
"Monty Python and the Holy Grail" PYTHON (MONTY) PICTURES LTD
/// Bram Moolenaar -- Br...@Moolenaar.net -- http://www.Moolenaar.net \\\
/// sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\ download, build and distribute -- http://www.A-A-P.org ///
\\\ help me help AIDS victims -- http://ICCF-Holland.org ///
Andy Wokula
> Alex wrote:
>> Hi all,
>>
>> I understand that these bugs and feature requests are "nice to have",
>> but anyway...
>>
>> VIM is not working very well with NTFS streams. When you edit file,
>> say, "test.txt:aaa",
>> it complains about backup during save (E510). If you edit file
>> "test.txt:my.txt", VIM opens
>> and edits file test_txt (but with correct stream ).
Just fooled around with this (I don't know much about NTFS streams ...).
Things to try out when writing the stream:
(1) simply add a bang:
:w!
(2) unset 'writebackup'
:h 'wb
:set nowb
:w
(3) set 'backupcopy' to "yes" (default is "auto")
:h 'bkc
:set wb bkc=yes
:w
...
To me, (3) looks ok, (2) and (1) less ok ...
I'd set these options only temporarily, when needed.
There must be some explicit support for streams, e.g. the swap file
is created as another stream.
--
Andy
krischik
wrote:
> 2 notes here:
> 1. I think NTFS streams is useless feature.
MacOS or extended attributes on OS/2 / Linux are used. And in fact
they are used for precisely that. Also note that extended attributes
on NTFS are implemented as streams as well. And Unix file attributes
(at least on Windows Server).
In fact I very much miss all the features OS/2 had to offer by the use
of extended attributes. For example I could attach a custom made icon
to important files (like cmd files) which would then be stored in the
extended attributes.
Or Rexx would store a pre-compiled version of all scripts inside the
extended attributes. And here streams would even excel since extended
attributes are limited in size while streams are not.
> I've never seen any
> practical example of the usefulness of this.
www.jpsoft.com/) and then use "DIR /:" a bit - you might be surprised
how many streams are already used on your system. Most notably:
12.12.2007 14:03 1'670 _H_A_________ .vimrc
0
{4c8cc155-6c1e-11d1-8e41-00c04fb9386d}:$DATA
Regards
Martin
George V. Reilly
On 31 Jan., 19:57, Alexei Alexandrov <alexei.alexand...@gmail.com>
wrote:
> 2 notes here:
> 1. I think NTFS streams is useless feature.
NTFS Streams can be used for the very same stuff resource streams on
MacOS or extended attributes on OS/2 / Linux are used. And in fact
But this is just about the only things that NTFS streams are ever used for. I spent ten years working in the Windows division at Microsoft, and I hardly ever saw streams being used.
> I've never seen any
> practical example of the usefulness of this.
You might want to get yourself the demo version of 4NT (http://
www.jpsoft.com/) and then use "DIR /:" a bit - you might be surprised
how many streams are already used on your system. Most notably:
12.12.2007 14:03 1'670 _H_A_________ .vimrc
0
{4c8cc155-6c1e-11d1-8e41-00c04fb9386d}:$DATA
Every NTFS file is implemented as one or more streams, and the default stream is called $DATA. It's a bit like finding a directory named . in every directory.
--
/George V. Reilly
http://www.georgevreilly.com/blog
krischik
On 1 Feb., 09:15, "George V. Reilly" <geo...@reilly.org> wrote:
> > You might want to get yourself the demo version of 4NT (http://
> >www.jpsoft.com/) and then use "DIR /:" a bit - you might be surprised
> > how many streams are already used on your system. Most notably:
>
> > 12.12.2007 14:03 1'670 _H_A_________ .vimrc
> > 0
> > {4c8cc155-6c1e-11d1-8e41-00c04fb9386d}:$DATA
>
> Every NTFS file is implemented as one or more streams, and the default
> stream is called $DATA. It's a bit like finding a directory named . in every
> directory.
C:\Documents and Settings\...\.vimrc
C:\Documents and Settings\...\.vimrc:
{4c8cc155-6c1e-11d1-8e41-00c04fb9386d}
The stream is called "{4c8cc155-6c1e-11d1-8e41-00c04fb9386d}". But
speaking of the directory "." - How about:
>dir /:
Datenträger in Laufwerk C ist SYSTEM Seriennummer ist
5cb9:81be
Verzeichnis von C:\Media\*
11.10.2007 12:28 <DIR> .
11.10.2007 12:28 <DIR> ..
646 $MountMgrRemoteDatabase:$DATA
28.09.2007 14:43 <JUNCTION> MEDIASAVE
[Volume{543aeb10-6dbe-11dc-b33a-001a73660515}\]
11.10.2007 12:28 <JUNCTION> TITANIUM [Volume{95a63aac-7702-11dc-
b340-001a73660515}\]
646 Bytes in 0 Dateien und 4 Verzeichnisse
28'062'420'992 Bytes frei
or
>dir /: /F
C:\Media\.
C:\Media\..
C:\Media\..:$MountMgrRemoteDatabase
C:\Media\MEDIASAVE
C:\Media\TITANIUM
Yep - a tream on directory ".." ;-).
Of course for all the "DIR" commands mentioned you need a command
interpreter which actually supports streams.
Martin
Alexei Alexandrov
> On 31 Jan., 19:57, Alexei Alexandrov <alexei.alexand...@gmail.com>
> wrote:
>
>> 2 notes here:
>
>> 1. I think NTFS streams is useless feature.
>
> NTFS Streams can be used for the very same stuff resource streams on
> MacOS or extended attributes on OS/2 / Linux are used.
I'm not aware well of Linux mechanisms for storing extended attributes
inside files on Linux. Word "Linux" looks strange to me here in fact -
it might a property of extX filesystem, but it doesn't have to do
anything with Linux I think since I may use other systems for root mount
point - reiserfs, for example. Do you have exact answers to the
following questions:
* Which mechanism is used on Linux to store these extended attributes on
Linux? Is it filesystem-based? If yes, which filesystems currently have
it implemented?
* Which applications use this feature on Linux? Are they OS-specific?
Are there application programs that use this? Are those programs
portable? If yes, how they deal with absence of this feature on other
file systems?
These are exactly questions which I would ask myself if I would be
designing something new and would be considering using NTFS file
streams. Will there be FAT32 file system clients? Would I care to port
the program to other system some time later? Et cetera. And most (if not
all) programs choose to use simple abstractions (files/directories)
available on all modern systems because it works well, because you'd
better keep it simple and because there are more important things to
focus on.
>
>> I've never seen any
>> practical example of the usefulness of this.
>
> You might want to get yourself the demo version of 4NT (http://
> www.jpsoft.com/) and then use "DIR /:" a bit - you might be surprised
> how many streams are already used on your system. Most notably:
>
> 12.12.2007 14:03 1'670 _H_A_________ .vimrc
> 0
> {4c8cc155-6c1e-11d1-8e41-00c04fb9386d}:$DATA
>
This looks like a default data stream in this file. I don't see that
there are 2 streams here - am I overlooking something?
P.S. I do see some cases where the data streams feature might be useful.
For example, anti-virus program might store some information about the
scanned file in a separate stream. But when I think of it more I realize
that even in these cases data streams approach would be questionable and
that there other ways to implement it with potentially better
performance and less limitations. For example, you cannot attach a new
data stream to a file which is available to you as read-only.
P.P.S. Thanks for reading to this point! :) Sorry for a long post and
somewhat clumsy English.
--
Alexei Alexandrov
krischik
wrote:
> > On 31 Jan., 19:57, Alexei Alexandrov <alexei.alexand...@gmail.com>
> > wrote:
>
> >> 2 notes here:
>
> >> 1. I think NTFS streams is useless feature.
>
> > NTFS Streams can be used for the very same stuff resource streams on
> > MacOS or extended attributes on OS/2 / Linux are used.
>
> I'm not aware well of Linux mechanisms for storing extended attributes
> inside files on Linux.
http://linux.die.net/man/5/attr
> Word "Linux" looks strange to me here in fact -
> it might a property of extX filesystem, but it doesn't have to do
> anything with Linux I think since I may use other systems for root mount
> point - reiserfs, for example.
support extended attributes as well. See:
http://en.wikipedia.org/wiki/Extended_file_attributes
> Do you have exact answers to the
> following questions:
> * Which mechanism is used on Linux to store these extended attributes on
> Linux? Is it filesystem-based? If yes, which filesystems currently have
> it implemented?
XFS filesystems support extended attributes (abbreviated xattr) if the
libattr feature is enabled in the kernel configuration. "
Which it is ever since about SuSE 9.0
> * Which applications use this feature on Linux? Are they OS-specific?
are not OS specific - they are supported via the "libattr" library.
> Are there application programs that use this? Are those programs
> portable? If yes, how they deal with absence of this feature on other
> file systems? These are exactly questions which I would ask myself if I would be
> designing something new and would be considering using NTFS file
> streams.
engineering for the last 20 years or so.
> Will there be FAT32 file system clients? Would I care to port
> the program to other system some time later? Et cetera. And most (if not
> all) programs choose to use simple abstractions (files/directories)
> available on all modern systems because it works well, because you'd
> better keep it simple and because there are more important things to
> focus on.
with 5 lines of code where solving the same problem without could cost
you 50. Determine file types, text file line endings and text file
encoding come to my mind here. Ask Bram how many line of code he
needed in Vim to determine these three informations. With consequent
use of xattribs it would have been 6 lines:
char mime_type[64];
char mime_encoding[64];
char line_ending[64];
getxattr (filename, "Content-Type",mime_type, sizeof mime_type);
getxattr (filename, "Line-Ending", line_ending, sizeof line_ending);
getxattr (filename, "Content-Encoding",mime_encoding, sizeof
mime_encoding);
And best of all: you know before you open the file. AFAIK Bram need to
close and reopen files in unfortunate combinations.
It is often overlooked that not only using a new feature posses risks
but not using as well.
How often I ended up with xxx.yyy.htm after a download... Or ever
better: the amount of frustrated SymbianOS users which tried to
download a theme (named xxx.utz) and ended up with xxx.zip. All
because some webserver and or browser made a mistake when trying to
determining the file type (utz and zip share the same file header).
And it is precisely those "simple abstractions (files/directories)
available on all modern systems" which bring those problems. Only:
It's not "modern systems" - it's "between 1 and 2 decades old
systems". Truly modern operating systems support extended attributes.
> P.S. I do see some cases where the data streams feature might be useful.
> For example, anti-virus program might store some information about the
> scanned file in a separate stream.
Martin
krischik
> And it is precisely those "simple abstractions (files/directories)
> available on all modern systems" which bring those problems. Only:
> It's not "modern systems" - it's "between 1 and 2 decades old
> systems". Truly modern operating systems support extended attributes.
that I used OS/2 before where storing the file type in the extended
attributes was standart. And it worked a lot better then extensions
(Dos) or scanning the first 256 bytes for magic patterns (Unix). There
was no guessing involved: The first filetype (the .TYPE xattrib in OS/
2 is a list) was the primary filetype the rest where possible
alternative representations (like a html file could also be
represented as plain text).
And then there was the .ICON xattrib - how I miss the .ICON xattrib -
I could drag and drop an icon to every file and from thereon the file
would be represented by that icon.
Now, nostalgia besides: The point is that those "simple abstractions"
lead to primitive solutions. And indeed OS/2 with it's advanced file
system offered a user experience which even Vista won't offer you. KDE
offers some of it - for the price of littering you filesystem with
".directory" files. But KDE still can't do what OS/2 could do - mainly
because they would need use extended attributes to do so - and the KDE
developers follow the same line of thought you do - which in turn
holds them back.
Martin
PS: On FAT the extended attributes where stored in a file called "EA
DATA. SF" - which shows that even older file systems can be taught new
tricks.
Matt Wozniski
>
> On the other hand use of extended attributes could solve a problem
> with 5 lines of code where solving the same problem without could cost
> you 50. Determine file types, text file line endings and text file
> encoding come to my mind here. Ask Bram how many line of code he
> needed in Vim to determine these three informations. With consequent
> use of xattribs it would have been 6 lines:
>
> char mime_type[64];
> char mime_encoding[64];
> char line_ending[64];
>
> getxattr (filename, "Content-Type",mime_type, sizeof mime_type);
> getxattr (filename, "Line-Ending", line_ending, sizeof line_ending);
> getxattr (filename, "Content-Encoding",mime_encoding, sizeof
> mime_encoding);
While this would be nice, it would require support code from every
application you have. It may only be 6 lines, but 6 lines * 5000
binaries is much more code than is in vim for line ending detection.
After all, vim can't use this information unless something put it
there, which requires that everything that can create a file - from
touch to wget to shell redirection - needs to be able to put that
attribute into the file. And, AFAIK, there's no way for those wget or
shell redirection to even know what type of line ending the data that
they wrote out had. Which means they'd need to detect it. Which
would require that just about every program out there that is
transcribing a data stream from one spot to another, rather than
authoring it itself, would need to duplicate the sort of EOL detection
code in vim in itself. And vim would still need to continue
supporting the old way since it's designed to run on filesystems that
still don't support extended attributes - or it would need to come up
with a way to fake extended attributes on those FS types. No one is
saying that extended attributes can't do useful things - just that
without being in any way standardized, it's unreasonable to expect
that any attributes will ever be set by anyone but you.
That being said, if vim had a convenient way to see if extended
attributes support was available on the filesystem, it might be worth
caching things like the fileformat in the extended attributes, so it
only needed to be computed once... Though that might open up a whole
new can of worms, since someone could easily change the line endings
with another tool between two vim runs, confusing the n00b user with a
bunch of ^M's show up everywhere... all in all, this is one place
where manual detection seems the best idea.
For things like content encoding, it might be more useful - but again,
since every other tool doesn't create it, vim would still need to fall
back on autodetection if it couldn't find an attribute for it.
~Matt
krischik
On 4 Feb., 21:10, "Matt Wozniski" <m...@drexel.edu> wrote:
> On Feb 4, 2008 9:18 AM, krischik wrote:
>
>
>
> > On the other hand use of extended attributes could solve a problem
> > with 5 lines of code where solving the same problem without could cost
> > you 50. Determine file types, text file line endings and text file
> > encoding come to my mind here. Ask Bram how many line of code he
> > needed in Vim to determine these three informations. With consequent
> > use of xattribs it would have been 6 lines:
>
> > char mime_type[64];
> > char mime_encoding[64];
> > char line_ending[64];
>
> > getxattr (filename, "Content-Type",mime_type, sizeof mime_type);
> > getxattr (filename, "Line-Ending", line_ending, sizeof line_ending);
> > getxattr (filename, "Content-Encoding",mime_encoding, sizeof
> > mime_encoding);
>
> While this would be nice, it would require support code from every
> application you have. It may only be 6 lines, but 6 lines * 5000
> binaries is much more code than is in vim for line ending detection.
webserver and browser getting detection wrong. konquror/nautilus/
explorer needs detection and far more complex then vim. And sure OS/2
needed detection. But there you could always overwrite faulty detected
value by manualy correcting the EA's.
> After all, vim can't use this information unless something put it
> there, which requires that everything that can create a file - from
> touch to wget to shell redirection - needs to be able to put that
> attribute into the file. And, AFAIK, there's no way for those wget or
> shell redirection to even know what type of line ending the data that
> they wrote out had. Which means they'd need to detect it.
attributes. Once it catches on then more application will support it.
But I know that this won't happen. We went down the "worse is better"
way for far to long.
> Which
> would require that just about every program out there that is
> transcribing a data stream from one spot to another, rather than
> authoring it itself, would need to duplicate the sort of EOL detection
> code in vim in itself. And vim would still need to continue
> supporting the old way since it's designed to run on filesystems that
> still don't support extended attributes - or it would need to come up
> with a way to fake extended attributes on those FS types. No one is
> saying that extended attributes can't do useful things - just that
> without being in any way standardized, it's unreasonable to expect
> that any attributes will ever be set by anyone but you.
advanced as much in the last 15 years as one would have expected
seeing the 15 years before: Backward compatibilty has hobbled us.
Martin
Matt Wozniski
>
> On 4 Feb., 21:10, "Matt Wozniski" wrote:
> >
> > While this would be nice, it would require support code from every
> > application you have. It may only be 6 lines, but 6 lines * 5000
> > binaries is much more code than is in vim for line ending detection.
>
> But the other 5000 applications need detection as well. I did mention
> webserver and browser getting detection wrong. konquror/nautilus/
> explorer needs detection and far more complex then vim. And sure OS/2
> needed detection. But there you could always overwrite faulty detected
> value by manualy correcting the EA's.
I'm not sure that most other apps do need detection. wget, for
instance, doesn't have to care what line endings the data it saves
has. But, for what you want, it would have to detect and save it,
even though it doesn't use it. Shell redirection has no idea what's
going through it, so it has no way to possibly say what Content-type
it is... or dd... or tar extracting files that were created on a
filesystem that didn't track extended attributes...
> > After all, vim can't use this information unless something put it
> > there, which requires that everything that can create a file - from
> > touch to wget to shell redirection - needs to be able to put that
> > attribute into the file. And, AFAIK, there's no way for those wget or
> > shell redirection to even know what type of line ending the data that
> > they wrote out had. Which means they'd need to detect it.
>
> Just one example: The OS/2 version of ZIP would pack extended
> attributes. Once it catches on then more application will support it.
> But I know that this won't happen. We went down the "worse is better"
> way for far to long.
The issue isn't waiting for it to catch on, it's that until it's
universally available it only increases the amount of code and
maintenance burden. And, there's no way that it would ever be
universally available for things like line endings style, because most
applications just don't know or need to care, so have no reason to set
that attribute. Things like marking the content encoding might be
more useful, but still, not good to work with... If, for example, you
had a file "test.txt" encoded in UTF-16 and with extended attributes
marking it as such, and your locale is set to use UTF-8, what would
you expect the result of "cp test.txt test2.txt" to be? What about
"cat text.txt >text2.txt"? If you expect those two commands to have
the same effect, I don't see how it can be done without changes to cp
(mark attrs on dest), cat (mark attrs on stdout), and the kernel
itself (allow extended attributes on streams).
> > Which
> > would require that just about every program out there that is
> > transcribing a data stream from one spot to another, rather than
> > authoring it itself, would need to duplicate the sort of EOL detection
> > code in vim in itself. And vim would still need to continue
> > supporting the old way since it's designed to run on filesystems that
> > still don't support extended attributes - or it would need to come up
> > with a way to fake extended attributes on those FS types. No one is
> > saying that extended attributes can't do useful things - just that
> > without being in any way standardized, it's unreasonable to expect
> > that any attributes will ever be set by anyone but you.
>
> Ahh, indeed, there is the problem and the reason why computing has not
> advanced as much in the last 15 years as one would have expected
> seeing the 15 years before: Backward compatibilty has hobbled us.
The problem with this idea isn't the limitations inherent in backwards
compatibility - if it were worthwhile to do, it's no problem to have
code that uses attributes if possible and the old, backwards
compatible method if not. The problem is that creating this standard
and tying it properly into a POSIX OS would require a lot of low-level
changes. If we can agree on the requirements - that there must be
some number of predefined attributes that are set on every regular
file (and what about special files?), that the operating system must
handle initializing these attributes correctly when a file is created,
that streams would also need to have attributes, and that copying,
extracting, and in general "moving around" data will "just work", can
you see how difficult this would be to implement? To fully implement
this, the very least that's required is an operating system that only
supports filesystems that have some way of storing extended
attributes, a kernel that's able to initialize each of those
attributes on every file created, a shell that knows how to get/set
these attributes (and which it needs to set), and a copy of the
coreutils that are aware of the changes (so that "cp file1 file2"
creates file2 with the attrs of file1, not the defaults), at the very
least? And that work is just the baseline - the point where other
software can start counting on a base set of attributes that will
always exist, and trusting that they're correct. On a quick glance, I
see absolutely no way to do this properly without support from the OS.
~Matt
Ben Schmidt
> advanced as much in the last 15 years as one would have expected
> seeing the 15 years before: Backward compatibilty has hobbled us.
Indeed, I think it has. The Mac OS used to use resource forks and type attributes
which were beautiful. Now we have descended to the level of extensions and magic
numbers like everyone else. The filesystem supports resource forks and named
forks, but nobody's really using them yet--indeed, if anything, at the moment they
add confusion because some parts of the OS detect file types that way rather than
by extension, when available, so it's hard to know what's going on.
Ben.
Send instant messages to your online friends http://au.messenger.yahoo.com
Ben Schmidt
>> Ahh, indeed, there is the problem and the reason why computing has not
>> advanced as much in the last 15 years as one would have expected
>> seeing the 15 years before: Backward compatibilty has hobbled us.
>
> Indeed, I think it has. The Mac OS used to use resource forks and type attributes
> which were beautiful. Now we have descended to the level of extensions and magic
> numbers like everyone else. The filesystem supports resource forks and named
> forks, but nobody's really using them yet--indeed, if anything, at the moment they
> add confusion because some parts of the OS detect file types that way rather than
> by extension, when available, so it's hard to know what's going on.
And that sad story told, I do agree with the other posters who don't think this
has much use in Vim. I think to have some kind of consistency, probably an OS or
other fairly central component needs to lead the way, not a text editor.
krischik
> On Feb 4, 2008 4:29 PM, krischik wrote:
>
> > On 4 Feb., 21:10, "Matt Wozniski" wrote:
>
> > > While this would be nice, it would require support code from every
> > > application you have. It may only be 6 lines, but 6 lines * 5000
> > > binaries is much more code than is in vim for line ending detection.
>
> > But the other 5000 applications need detection as well. I did mention
> > webserver and browser getting detection wrong. konquror/nautilus/
> > explorer needs detection and far more complex then vim. And sure OS/2
> > needed detection. But there you could always overwrite faulty detected
> > value by manualy correcting the EA's.
>
> I'm not sure that most other apps do need detection. wget, for
> instance, doesn't have to care what line endings the data it saves
> has. But, for what you want, it would have to detect and save it,
> even though it doesn't use it.
file from an EA aware ftp server - the same way it request file
permissions today when used with "--preserve-permissions".
> Shell redirection has no idea what's
> going through it, so it has no way to possibly say what Content-type
> it is... or dd...
> or tar extracting files that were created on a
> filesystem that didn't track extended attributes...
been added to tar last year:
http://www.redhatmagazine.com/2007/07/02/tips-from-an-rhce-tar-vs-star-the-battle-of-xattrs/
> The issue isn't waiting for it to catch on, it's that until it's
> universally available it only increases the amount of code and
> maintenance burden.
development.
> Things like marking the content encoding might be
> more useful, but still, not good to work with... If, for example, you
> had a file "test.txt" encoded in UTF-16 and with extended attributes
> marking it as such, and your locale is set to use UTF-8, what would
> you expect the result of "cp test.txt test2.txt" to be?
are copied as well. At least on SuSE Linux 9.2 onwards. And of course
the file would still be UTF-16 - after all it's "copy" not "convert".
Without meaning offence: You should read up a little on the subject as
you knowledge is not up to date.
> What about
> "cat text.txt >text2.txt"? If you expect those two commands to have
> the same effect, I don't see how it can be done without changes to cp
> (mark attrs on dest), cat (mark attrs on stdout), and the kernel
> itself (allow extended attributes on streams).
that they never have behaved the same: cp has options like "--
preserve" and "--archive" - cat has not. Even today using cat to copy
a file will mean that you loose the all the meta informations
attached.
> a shell that knows how to get/set
> these attributes (and which it needs to set),
> and a copy of the
> coreutils that are aware of the changes (so that "cp file1 file2"
> creates file2 with the attrs of file1, not the defaults),
> On a quick glance, I
> see absolutely no way to do this properly without support from the OS.
here. Only I have once used an operating system (OS/2) which did
exactly that - hence my frustration.
Martin
krischik
> And that sad story told, I do agree with the other posters who don't think this
> has much use in Vim. I think to have some kind of consistency, probably an OS or
> other fairly central component needs to lead the way, not a text editor.
forks" are useless. And I only used Vim as an example of where Vim
would have been simpler to implement - on an operating system which
properly supports "type attributes".
Sad that everybody else only sees only hurdles and risks of change and
not the chances and the risk of no-change.
And yes: there is a rist of no-change!
The risk of an ever more complex software world with ever more complex
solutions which could be done much easier if one would have done it
right from the beginning.
And the risk of ever more exploits of those complex solutions.
Martin
Tony Mechelynck
[...]
> I would use "cp --archive test.txt test2.txt" in which case the EA's
> are copied as well. At least on SuSE Linux 9.2 onwards. And of course
> the file would still be UTF-16 - after all it's "copy" not "convert".
>
> Without meaning offence: You should read up a little on the subject as
> you knowledge is not up to date.
> Done: GNU cp will do that if --archive is used.
Hmm... According to "info cp" (on my openSUSE 10.3 system), -a or --archive is
equivalent to -dpPR, which means:
-d copy symlinks as symlinks and preserve hardlinks between sources in the copies
-p preserve attributes. If not specifying which attributes, the default is:
mode,ownership,timestamps (xattrs must be specified explicitly to be included)
-P copy symlinks as symlinks (sic)
-R copy directories recursively
Best regards,
Tony.
--
"She said, `I know you ... you cannot sing'. I said, `That's nothing,
you should hear me play piano.'"
-- Morrisey
krischik
On 5 Feb., 11:37, Tony Mechelynck <antoine.mechely...@gmail.com>
wrote:
>
> [...]> I would use "cp --archive test.txt test2.txt" in which case the EA's
> > are copied as well. At least on SuSE Linux 9.2 onwards. And of course
> > the file would still be UTF-16 - after all it's "copy" not "convert".
>
> > Without meaning offence: You should read up a little on the subject as
> > you knowledge is not up to date.
> [...]
> > Done: GNU cp will do that if --archive is used.
>
> [...]
>
> Hmm... According to "info cp" (on my openSUSE 10.3 system), -a or --archive is
> equivalent to -dpPR, which means:
>
> -d copy symlinks as symlinks and preserve hardlinks between sources in the copies
> -p preserve attributes. If not specifying which attributes, the default is:
> mode,ownership,timestamps (xattrs must be specified explicitly to be included)
> -P copy symlinks as symlinks (sic)
> -R copy directories recursively
creates them at mass) and saw what happened. But that was on a fairly
old SuSE 9.2. I check SuSE 10.3 when I am home.
Martin
Paul LeoNerd Evans
krischik <kris...@users.sourceforge.net> wrote:
> Sad that everybody else only sees only hurdles and risks of change and
> not the chances and the risk of no-change.
>
> And yes: there is a rist of no-change!
I have to agree with krischik here. At University, I once got stuck in a
problem with some of the other mathematicians, who were all of the
opinion that "Yes, I'll have tea if you're having tea". Problem was,
nobody actually had tea. It took one of the lawyers to stand up and say
"Alright then, I'll have tea." Then all the mathmos jumped ship.
An amusing story perhaps, but it proves the point. In all things like
this, someone needs to jump first. Does it really matter who that is, as
long as someone does?
It's been a bone of contention of mine for years, the way modified keys
work in terminals. Five years ago XTerm gained ways to express generic
modified keys (e.g. Ctrl-Shift-Left). Only recently did any application,
such as Vim, start to understand those. It's always been a "you jump,
I'll jump" situation - why should the terminal send sequences nobody
would understand, or why should any application look for sequences nobody
would send? Someone has to go first.
Back to the subject of EAs - someone already has gone first. The GNU
fileutils already support EAs. As does tar. I also know that the lighttpd
webserver uses them by default, only falling back on filename-based
detection if the file doesn't have a "Content-Type" EA.
It'd be lovely if Vim could set those for it.
--
Paul "LeoNerd" Evans
leo...@leonerd.org.uk
ICQ# 4135350 | Registered Linux# 179460
http://www.leonerd.org.uk/
krischik
> I also know that the lighttpd
> webserver uses them by default, only falling back on filename-based
> detection if the file doesn't have a "Content-Type" EA.
research before I rant) I found out that Apache too uses EA's when to
store and detect file types.
And then there is Samba which uses "user.DOSATTRIB" to store dos
attributes.
> It'd be lovely if Vim could set those for it.
Martin
Alexei Alexandrov
>
> Open a console and type "man 5 attr"? Or look here:
>
> http://linux.die.net/man/5/attr
>
>
> support extended attributes as well. See:
>
> http://en.wikipedia.org/wiki/Extended_file_attributes
>
> XFS filesystems support extended attributes (abbreviated xattr) if the
> libattr feature is enabled in the kernel configuration. "
Oh, thanks - this is a good learning for me. It's a shame I never heard
of this before.
>
> On the other hand use of extended attributes could solve a problem
> with 5 lines of code where solving the same problem without could cost
> you 50. Determine file types, text file line endings and text file
> encoding come to my mind here. Ask Bram how many line of code he
> needed in Vim to determine these three informations. With consequent
> use of xattribs it would have been 6 lines:
>
> [...]
>
> And best of all: you know before you open the file. AFAIK Bram need to
> close and reopen files in unfortunate combinations.
>
Such attributes would be useful, yes. But it's backward compatibility
that is the root of all evil and you also mentioned this in a sibling
post. So I'm not sure these things will become popular unless there is a
brand new OS appears which doesn't have to be compatible with anything
and it wins the market.
Also, the named attributes story doesn't have to do much with numbered
NTFS data streams, does it? It might be similar but numbers are too far
from string identifier convenience. Which brings me back to my original
thought - the design and implementation of data streams in NTFS are useless.
--
Alexei Alexandrov
krischik
wrote:
> Also, the named attributes story doesn't have to do much with numbered
> NTFS data streams, does it?
supported. They are called "Alternate data streams". See
http://en.wikipedia.org/wiki/NTFS#Features
Martin
Bram Moolenaar
Alex Jakushev wrote:
> On 1/31/08, Bram Moolenaar <Br...@moolenaar.net> wrote:
> >
> > Try setting the 'backupcopy' option to "yes".
>
> This worked, but then it raises another question. Previous value
> of the 'backupcopy' option was auto, which means yes or no, which
> works best. Why didn't it choose yes, if no fails?
Vim doesn't detect the situation that you are editing an info stream
directly.
> Also, in my configuration, backup option is no. Theoretically, I should
> not bother with all this stuff at all? I use Vim7.1 last official release.
The 'writebackup' option matters too.
> > I thought Vim did copy the streams, but I suppose this doesn't work when
> > you edit one specific stream. If you want to look at it: function
> > copy_infostreams() in src/os_win32.c
>
> This may take a while for me to set up everything...
>
> But i have an idea that the problem is with file names.
> When you open some specific stream, vim considers the stream name as
> filename, and filename as subfolder. The stream name is separated from file
> name with a colon (c:\path\foo.txt:bar), and the same colon is used to
> separate drive letter. Maybe this causes some confusion?
Detecting the colon should be simple. When not using NTFS the file name
would be illegal. I'll add a todo item for this. But it would be nice
if someone can make a patch for it.
> Also, VIM converts dots in file name to underscores sometimes, when
> streams are specified.
I haven't seen this. Perhaps it's because the long file name is
converted to a 8.3 file name? Can you give a reproducable example?
> > NTFS streams are mostly restricted to MS-Windows applications, and
> > rarely used. I don't think it's a good idea to support them in Vim
> > directly.
>
> I understand it, but vim has specific windows functionality anyway.
> But of course it is easy for me to suggest this and that :)
I always put more effort in functionality that works everywhere. And to
keep programs portable rare features should be avoided, especially when
they don't add something essential for the end user. MS thinks
otherwise: They want programs to only run on MS-Windows. Portability
means they might lose customers. Some Linux developers also go in this
direction, they don't care about MS-Windows users. I care for everybody
:-).
--
The process for understanding customers primarily involves sitting around with
other marketing people and talking about what you would to if you were dumb
enough to be a customer.
(Scott Adams - The Dilbert principle)
Alexei Alexandrov
>
> I fear you missed something - since NTFS v3.0 named forks are
> supported. They are called "Alternate data streams".
>
Oops, yep, I have a lot of stuff to catch up with.
--
Alexei Alexandrov