[9fans] file creation time

[9fans] file creation time

[lucio]

Does Plan 9 really not provide for file creation?

	typedef
	struct Dir {
		/* system-modified data */
		ushort	type;	/* server type */
		uint	dev;	/* server subtype */
		/* file data */
		Qid	qid;	/* unique id from server */
		ulong	mode;	/* permissions */
		ulong	atime;	/* last read time */
		ulong	mtime;	/* last write time */
		vlong	length;	/* file length */
		char	*name;	/* last element of path */
		char	*uid;	/* owner name */
		char	*gid;	/* group name */
		char	*muid;	/* last modifier name */
	} Dir;

It's been a long time, have I asked this already?

++L

Re: [9fans] file creation time

[erik quanstrom]

On Thu Jan 22 12:41:10 EST 2009, lucio@proxima.alt.za wrote:
> Does Plan 9 really not provide for file creation?
>
> 	typedef
> 	struct Dir {
> 		/* system-modified data */
> 		ushort	type;	/* server type */
> 		uint	dev;	/* server subtype */
> 		/* file data */
> 		Qid	qid;	/* unique id from server */
> 		ulong	mode;	/* permissions */
> 		ulong	atime;	/* last read time */
> 		ulong	mtime;	/* last write time */
> 		vlong	length;	/* file length */
> 		char	*name;	/* last element of path */
> 		char	*uid;	/* owner name */
> 		char	*gid;	/* group name */
> 		char	*muid;	/* last modifier name */
> 	} Dir;
>
> It's been a long time, have I asked this already?

use the dump, luke. ☺

- erik

Re: [9fans] file creation time

[lucio]

> use the dump, luke. ☺

If there was an easy, foolproof way to scan the dump by filename, I
presume I could search for the earliest instance and consider that the
time of creation.  Not entirely viable, is it?

I do wonder why this field was sacrificed in the file system(s) and
9P*?

Or am I dreaming and the time of creation is just a figment of my
imagination?

++L

PS: I'm sure your reply was in jest.  So, largely, is my query.

Re: [9fans] file creation time

[andrey mirtchovski]

On Thu, Jan 22, 2009 at 10:50 AM,  <lucio@proxima.alt.za> wrote:
>> use the dump, luke. ☺
>
> If there was an easy, foolproof way to scan the dump by filename, I
> presume I could search for the earliest instance and consider that the
> time of creation.  Not entirely viable, is it?
>

history(1)

http://plan9.bell-labs.com/magic/man2html?man=history&sect=1

Re: [9fans] file creation time

[erik quanstrom]

On Thu Jan 22 12:54:14 EST 2009, lucio@proxima.alt.za wrote:
> > use the dump, luke. ☺
>
> If there was an easy, foolproof way to scan the dump by filename, I
> presume I could search for the earliest instance and consider that the
> time of creation.  Not entirely viable, is it?

it's hard to give a good answer, since i don't
know what the goal is.

> I do wonder why this field was sacrificed in the file system(s) and
> 9P*?
>
> Or am I dreaming and the time of creation is just a figment of my
> imagination?

of course on a worm filesystem, you could say that each
time a file is modified after it has made it to the worm, it
becomes a brand new file.  though the qid.path doesn't
track this view.  the create time would then differ
from the mtime by as most the time between dumps.

andrey suggests

> history(1)

history doesn't give creation times.  it gives the last mtime
before the first dump of the file.  the file could be deleted
and recreated many times and you'd never know.  (unless
you were to peek at the qid.path.)

but then again, history might be exactly what you're
looking for.  it usually is for me.

- erik

Re: [9fans] file creation time

[roger peppe]

2009/1/22  <lucio@proxima.alt.za>:
> I do wonder why this field was sacrificed in the file system(s) and
> 9P*?

useful or not, traditional unix ctime never gave the creation time
anyway - it gave inode modification time, which isn't the
same thing at all: it's updated when you do a chmod.

Re: [9fans] file creation time

[lucio]

> useful or not, traditional unix ctime never gave the creation time
> anyway - it gave inode modification time, which isn't the
> same thing at all: it's updated when you do a chmod.

Well, that is understandable, if contrary to the principle of least
astonishment.  Thank you for that gem, I would never have picked it
up without having it pointed out to me.

++L

Re: [9fans] file creation time

[Charles Forsyth]

>I would never have picked it up without having it pointed out to me.

to be fair, in one of the system include files it was commented as
ino.h:	time_t	di_ctime;   	/* time created */
but it was only a comment.  stat(2) was more accurate.

Re: [9fans] file creation time

[erik quanstrom]

On Thu Jan 22 13:17:59 EST 2009, rogpeppe@gmail.com wrote:
> 2009/1/22  <lucio@proxima.alt.za>:
> > I do wonder why this field was sacrificed in the file system(s) and
> > 9P*?
>
> useful or not, traditional unix ctime never gave the creation time
> anyway - it gave inode modification time, which isn't the
> same thing at all: it's updated when you do a chmod.

will the real creation time please stand up.

- erik

Re: [9fans] file creation time

[Richard Miller]

> will the real creation time please stand up.

October 23, 4004 BC.

Re: [9fans] file creation time

[erik quanstrom]

On Thu Jan 22 14:33:05 EST 2009, 9fans@hamnavoe.com wrote:
> > will the real creation time please stand up.
>
> October 23, 4004 BC.

so when we add 8-byte times to 9p2010 Tstat
and Wstat, we can add a constant ctime field of
-188466825600?  or should that just be hardcoded?

- erik

Re: [9fans] file creation time

[Charles Forsyth]

> Does Plan 9 really not provide for file creation?
no, but it wasn't really creation on unix but inode change.