anoncvs and diff

Nigel J. Andrews wrote:

I've been waiting to see how a patched file differs from my version.

The patch was added to the to apply list last week I think (it wasn't mine btw)
and I've been doing cvs diff to view the differences so I can tell when the
patch has been applied. Additional information given by this is the revision
number the comparison is against of course. This has stayed at 1.61 all the
time I've been doing this cvs diff operation. Looking at the web interface to
cvs I see the file has a revision number of 1.64. I use the anoncvs server for
my operations. Am I being daft or is there a problem with the anoncvs archive?

That is strange. anoncvs and the web interface should have the same
version number. What file are you looking at?

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

On Thu, 3 Oct 2002, Bruce Momjian wrote:

Nigel J. Andrews wrote:

I've been waiting to see how a patched file differs from my version.

The patch was added to the to apply list last week I think (it wasn't mine btw)
and I've been doing cvs diff to view the differences so I can tell when the
patch has been applied. Additional information given by this is the revision
number the comparison is against of course. This has stayed at 1.61 all the
time I've been doing this cvs diff operation. Looking at the web interface to
cvs I see the file has a revision number of 1.64. I use the anoncvs server for
my operations. Am I being daft or is there a problem with the anoncvs archive?

That is strange. anoncvs and the web interface should have the same
version number. What file are you looking at?

src/pl/tcl/pltcl.c

However, since writing that I've tried some other things.

cvs diff -r HEAD pltcl.c

gave me differences against revision 1.64

and cvs update pltcl.c

said it was merging changes between 1.64 and 1.61

and a plain cvs diff now shows me differences against 1.64

I think this is probably just a short fall in my fairly basic knowledge of how
cvs works.

--
Nigel J. Andrews

Nigel J. Andrews wrote:

cvs diff -r HEAD pltcl.c

gave me differences against revision 1.64

and cvs update pltcl.c

said it was merging changes between 1.64 and 1.61

and a plain cvs diff now shows me differences against 1.64

I think this is probably just a short fall in my fairly basic knowledge of how
cvs works.

What does 'cvs log' say about the file, especially the top stuff?

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

On Thu, 3 Oct 2002, Bruce Momjian wrote:

Nigel J. Andrews wrote:

cvs diff -r HEAD pltcl.c

gave me differences against revision 1.64

and cvs update pltcl.c

said it was merging changes between 1.64 and 1.61

and a plain cvs diff now shows me differences against 1.64

I think this is probably just a short fall in my fairly basic knowledge of how
cvs works.

What does 'cvs log' say about the file, especially the top stuff?

It gave me the log all the way up to the 1.64 revision with the REL7_3_STABLE
label assigned to revision 1.64.0.2

Revision 1.64 apparently backing out my patch which made 1.63.

I had a brain wave and did the cvs log command which was what lead me to try
specifying revisions. As I say it looks like a lack of knowledge about how cvs
works for these things. I always thought it worked like RCS and gave a diff
against the latest checked in but obviously not.

BTW, I've found Neil Conway's patch for this file, email dated 25th Sept., I
can forward it or apply it and include the changes along with whatever I do for
my next submission, which ever you'd prefer. I'd suggest it's easy to let me
apply and submit it due to overlaps.

--
Nigel J. Andrews

Nigel J. Andrews wrote:

It gave me the log all the way up to the 1.64 revision with the REL7_3_STABLE
label assigned to revision 1.64.0.2

Revision 1.64 apparently backing out my patch which made 1.63.

I had a brain wave and did the cvs log command which was what lead me to try
specifying revisions. As I say it looks like a lack of knowledge about how cvs
works for these things. I always thought it worked like RCS and gave a diff
against the latest checked in but obviously not.

BTW, I've found Neil Conway's patch for this file, email dated 25th Sept., I
can forward it or apply it and include the changes along with whatever I do for
my next submission, which ever you'd prefer. I'd suggest it's easy to let me
apply and submit it due to overlaps.

Sure, sounds good.

-- 
  Bruce Momjian                        |  http://candle.pha.pa.us
  pgman@candle.pha.pa.us               |  (610) 359-1001
  +  If your life is a hard drive,     |  13 Roberts Road
  +  Christ can be your backup.        |  Newtown Square, Pennsylvania 19073

"Nigel J. Andrews" <nandrews@investsystems.co.uk> writes:

I had a brain wave and did the cvs log command which was what lead me to try
specifying revisions. As I say it looks like a lack of knowledge about how cvs
works for these things. I always thought it worked like RCS and gave a diff
against the latest checked in but obviously not.

I think "cvs diff foo.c" without any switches gives you the diff between
your local copy of foo.c and the last version of foo.c *that you checked
out* --- ie, it shows you the uncommitted editing that you've done.

If you hadn't done "cvs update" since rev 1.61 then this would explain
the behavior you saw.

regards, tom lane