threads / bug / 7609

cvsserver bug

Subject: cvsserver bug

## tl;dr

9 messages between Apr 11, 2007 and Apr 11, 2007.

replies: 8people: 4as markdown or json

Daniel Barkalow· Apr 11, 2007, 16:10 UTC · lore

It seems like git-cvsserver doesn't know the CVS special case that, if the client has removed the file from the working directory (but not called "cvs remove"), this means to revert it to the server's version. I think that the condition around line 843 needs to exclude this case, and it needs to get to line 892 instead, but I can't even fake perl to fix it.

	-Daniel
*This .sig left intentionally blank*
Frank Lichtenheld· Apr 11, 2007, 19:43 UTC · re: Daniel Barkalow · lore

Re: cvsserver bug

On Wed, Apr 11, 2007 at 12:10:24PM -0400, Daniel Barkalow wrote:
Show 5 quoted lines
> It seems like git-cvsserver doesn't know the CVS special case that, if the 
> client has removed the file from the working directory (but not called 
> "cvs remove"), this means to revert it to the server's version. I think 
> that the condition around line 843 needs to exclude this case, and it 
> needs to get to line 892 instead, but I can't even fake perl to fix it.

I can confirm the bug. Will look into it a bit and see whether I can come up with a fix.

Gruesse,
-- 
Frank Lichtenheld <frank@lichtenheld.de>
www: http://www.djpig.de/
Frank Lichtenheld· Apr 11, 2007, 20:38 UTC · re: Daniel Barkalow · lore

[PATCH] cvsserver: Fix handling of diappeared files on update

Only send a modified response if the client sent a "Modified" entry. This fixes the case where the file was locally deleted on the client without being removed from CVS. In this case the client will only have sent the Entry for the file but nothing else.

Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de>
---
 git-cvsserver.perl |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
 We really, really need a test suite for cvsserver...
 I've tested this as good for regressions as I could
 think of but am still unsure about it.
diff --git a/git-cvsserver.perl b/git-cvsserver.perl
index 68aa752..25816c5 100755
--- a/git-cvsserver.perl
+++ b/git-cvsserver.perl
@@ -843,6 +843,7 @@ sub req_update
         if ( defined ( $wrev )
              and defined($meta->{revision})
              and $wrev == $meta->{revision}
+             and defined($state->{entries}{$filename}{modified_hash})
              and not exists ( $state->{opt}{C} ) )
         {
             $log->info("Tell the client the file is modified");
-- 
1.5.1
Junio C Hamano· Apr 11, 2007, 21:36 UTC · re: Frank Lichtenheld · lore

Re: [PATCH] cvsserver: Fix handling of diappeared files on update

Frank Lichtenheld <frank@lichtenheld.de> writes:
Show 30 quoted lines
> Only send a modified response if the client sent a
> "Modified" entry. This fixes the case where the
> file was locally deleted on the client without
> being removed from CVS. In this case the client
> will only have sent the Entry for the file but nothing
> else.
>
> Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de>
> ---
>  git-cvsserver.perl |    1 +
>  1 files changed, 1 insertions(+), 0 deletions(-)
>
>  We really, really need a test suite for cvsserver...
>  I've tested this as good for regressions as I could
>  think of but am still unsure about it.
>
> diff --git a/git-cvsserver.perl b/git-cvsserver.perl
> index 68aa752..25816c5 100755
> --- a/git-cvsserver.perl
> +++ b/git-cvsserver.perl
> @@ -843,6 +843,7 @@ sub req_update
>          if ( defined ( $wrev )
>               and defined($meta->{revision})
>               and $wrev == $meta->{revision}
> +             and defined($state->{entries}{$filename}{modified_hash})
>               and not exists ( $state->{opt}{C} ) )
>          {
>              $log->info("Tell the client the file is modified");
> -- 
> 1.5.1

This would make the modified response go away, but would it cause a fresh re-checkout to happen?

Daniel Barkalow· Apr 11, 2007, 21:39 UTC · re: Junio C Hamano · lore

Re: [PATCH] cvsserver: Fix handling of diappeared files on update

On Wed, 11 Apr 2007, Junio C Hamano wrote:
Show 35 quoted lines
> Frank Lichtenheld <frank@lichtenheld.de> writes:
> 
> > Only send a modified response if the client sent a
> > "Modified" entry. This fixes the case where the
> > file was locally deleted on the client without
> > being removed from CVS. In this case the client
> > will only have sent the Entry for the file but nothing
> > else.
> >
> > Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de>
> > ---
> >  git-cvsserver.perl |    1 +
> >  1 files changed, 1 insertions(+), 0 deletions(-)
> >
> >  We really, really need a test suite for cvsserver...
> >  I've tested this as good for regressions as I could
> >  think of but am still unsure about it.
> >
> > diff --git a/git-cvsserver.perl b/git-cvsserver.perl
> > index 68aa752..25816c5 100755
> > --- a/git-cvsserver.perl
> > +++ b/git-cvsserver.perl
> > @@ -843,6 +843,7 @@ sub req_update
> >          if ( defined ( $wrev )
> >               and defined($meta->{revision})
> >               and $wrev == $meta->{revision}
> > +             and defined($state->{entries}{$filename}{modified_hash})
> >               and not exists ( $state->{opt}{C} ) )
> >          {
> >              $log->info("Tell the client the file is modified");
> > -- 
> > 1.5.1
> 
> This would make the modified response go away, but would it
> cause a fresh re-checkout to happen?

It falls through into the appropriate case, evidently. At least, it seems to resolve my issue.

	-Daniel
*This .sig left intentionally blank*
Frank Lichtenheld· Apr 11, 2007, 21:42 UTC · re: Junio C Hamano · lore

Re: [PATCH] cvsserver: Fix handling of diappeared files on update

On Wed, Apr 11, 2007 at 02:36:10PM -0700, Junio C Hamano wrote:
> This would make the modified response go away, but would it
> cause a fresh re-checkout to happen?

Yeah, the rest of the code handles this case correctly if it actually gets the chance to do so.

Gruesse,
-- 
Frank Lichtenheld <frank@lichtenheld.de>
www: http://www.djpig.de/
Martin Langhoff· Apr 11, 2007, 21:52 UTC · re: Frank Lichtenheld · lore

Re: [PATCH] cvsserver: Fix handling of diappeared files on update

Frank Lichtenheld wrote:
Show 8 quoted lines
> Only send a modified response if the client sent a
> "Modified" entry. This fixes the case where the
> file was locally deleted on the client without
> being removed from CVS. In this case the client
> will only have sent the Entry for the file but nothing
> else.
> 
> Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de>

Ack. With this, git-cvsserver matches cvs's behaviour. The only difference is that cvs also spits out

  E cvs update: warning: $path was lost
cheers,
martin
-- 
-----------------------------------------------------------------------
Martin @ Catalyst .Net .NZ  Ltd, PO Box 11-053, Manners St,  Wellington
WEB: http://catalyst.net.nz/           PHYS: Level 2, 150-154 Willis St
OFFICE: +64(4)916-7224  UK: 0845 868 5733 ext 7224  MOB: +64(21)364-017
      Make things as simple as possible, but no simpler - Einstein
-----------------------------------------------------------------------
Daniel Barkalow· Apr 11, 2007, 22:02 UTC · re: Martin Langhoff · lore

Re: [PATCH] cvsserver: Fix handling of diappeared files on update

On Thu, 12 Apr 2007, Martin Langhoff wrote:
Show 14 quoted lines
> Frank Lichtenheld wrote:
> > Only send a modified response if the client sent a
> > "Modified" entry. This fixes the case where the
> > file was locally deleted on the client without
> > being removed from CVS. In this case the client
> > will only have sent the Entry for the file but nothing
> > else.
> > 
> > Signed-off-by: Frank Lichtenheld <frank@lichtenheld.de>
> 
> Ack. With this, git-cvsserver matches cvs's behaviour. The only
> difference is that cvs also spits out
> 
>   E cvs update: warning: $path was lost
My cvs client says:
cvs update: warning: $path unexpectedly disappeared

I'm not sure if something suitable is generated client-side in general, but mine seems to.

	-Daniel
*This .sig left intentionally blank*

← back to recent threads