bugArX revision control system - Bugs: bug #4760, faster what-changed if not making...

 
 

bug #4760: faster what-changed if not making patches

Submitted by:  Walter Landry <landry>
Submitted on:  Sat 16 Aug 2003 04:38:29 PM UTC  
 
Severity: 1 - FeatureItem Group: None
Status: Works For MePrivacy: Public
Assigned to: NoneOpen/Closed: Closed

Add a New Comment(Rich Markup)
   

You are not logged in

Please log in, so followups can be emailed to you.

 

Tue 30 Mar 2004 02:37:00 AM UTC, comment #1:

This kind of optimization has been already done for the longest time (1.0.3 I think). ArX would check whether two files are the same by reading in 64K blocks and comparing them. If there is any difference, then it does the full diff.

Walter Landry <landry>
Project Administrator
Sat 16 Aug 2003 04:38:29 PM UTC, original submission:

"Mark A. Flacy" <mflacy1@attbi.com> wrote:

>
> On 2003.03.10 15:08 Robert Anderson wrote:
> >
> > --- Original Message ---
> > From: Walter Landry <wlandry@ucsd.edu>
> > To: -unavailable-
> > Subject: Re: [arch-users] Re: BitBucket: GPL-ed KitBeeper clone
> >
> > >Bruce Stephens <bruce@cenderis.demon.co.uk> wrote:
> > >> "Robert Anderson" <RWA@sbcglobal.net> writes:
> > >>
> > >> [...]
> > >>
> > >> > Anyone know how cvs -n upd works, why it is so fast, and why
> > arch
> > >> > can't do something equivalent?
> > >>
> > >> I think it's doing an mtime check (using CVS/Entries). I'm
> > not quite
> > >> sure why that should be faster than what's in ArX.
> > >
> > >ArX still does a full diff (in addition to all of the shell
> > overhead).
> > >Doing ctime/mtime checks is something that has been endlessly
> > >discussed, but not yet implemented. It probably won't happen
> > until it
> > >is the absolutely last thing that can be done to improve
> > performance.
> >
> > I wonder if there isn't something in-between ctime/mtime checks
> > and diff -r.
>
> diff -r -q
>
>
> --
> Mark A. Flacy (972) 685-8347
> moc.ibtta@1ycalfm <<- To defeat not very bright spambots.


"Robert Anderson" <RWA@sbcglobal.net> wrote:

> >>"Robert Anderson" <RWA@sbcglobal.net> wrote:
> >>> I wonder if there isn't something in-between ctime/mtime checks
> >>> and diff -r.
> >>>
> >>> If a person is only interested in the "summary" information from
> >>> what-changed, might some kind of find/cmp solution be faster?
> >>
> >>I don't think that will help much. The main killer is having to
> >open
> >>2*N files, read their contents, and see if they differ. Actually
> >>writing the patches is probably not all that big of a slowdown.
> >
> >If that's what cmp does, it seems like a braindead implementation
> >on the surface. Why can't you prune all the files into the
> >"different" bin that simply have a different length as a first
> >pass? No opening or reading of contents or comparing needed.
>
> And on top of that, for files not already pruned, "cmp" can stop
> comparing at the first difference, whereas diff has to compare
> the full files, no?


You're right. I was thinking too much in the box. I'll file this as
a feature request. Right now, it probably won't make a huge
difference, but once everything is sped up it would be significant.

Walter

Walter Landry <landry>
Project Administrator

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach File(s):
   
   
Comment:
   

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

CC list is empty

 

Do you think this task is very important?
If so, you can click here to add your encouragement to it.
This task has 0 encouragements so far.

Only logged-in users can vote.

 

Please enter the title of George Orwell's famous dystopian book (it's a date):

 

 

Follow 2 latest changes.

Date Changed By Updated Field Previous Value => Replaced By
Tue 30 Mar 2004 02:37:00 AM UTClandryStatusNone=>Works For Me
  Open/ClosedOpen=>Closed

Back to the top


Powered by Savane 3.1-cleanup1