threads / discuss / 4189

cvsimport weird

Subject: cvsimport weird

## tl;dr

6 messages between May 18, 2006 and May 18, 2006.

replies: 5people: 4as markdown or json

Bertrand Jacquin· May 18, 2006, 01:00 UTC · lore

Hi,

I would like to make some git-cvsimport test on a public repo. And I get some problem and don't know if it's a remote server (enlightenment) problem or a git-cvsimport one.

Here is the log :

/mnt/data/src/e-tmp % git-cvsimport -v
-d:pserver:anonymous@anoncvs.enlightenment.org:/var/cvs/e e17
connect error: Network is unreachable
WARNING: malformed CVS version str: Server:
WARNING: Your CVS client version:
[Client: Concurrent Versions System (CVS) 1.12.12 (client)]
and/or server version:
[Server: ]
are too old to properly support the rlog command.
This command was introduced in 1.11.1.  Cvsps
will use log instead, but PatchSet numbering
may become unstable due to pruned empty
directories.

cvs [log aborted]: end of file from server (consult above messages if any) DONE. Already up-to-date. /mnt/data/src/e-tmp %

I use cvs 1.12.12, cvsps 2.1, git 1.3.3

I don't why it tell me that Network is unreachable as I can do normal cvs checkout.

Also, I have the repo in another directory, and I don't know how to use a :local: CVSROOT

-- 
Beber
#e.fr@freenode

Martin Langhoff· May 18, 2006, 02:54 UTC · re: Bertrand Jacquin · lore

Re: cvsimport weird

On 5/18/06, Bertrand Jacquin <beber.mailing@gmail.com> wrote:

Show 21 quoted lines

> Hi,
>
> I would like to make some git-cvsimport test on a public repo. And I
> get some problem and don't know if it's a remote server
> (enlightenment) problem or a git-cvsimport one.
>
> Here is the log :
>
> /mnt/data/src/e-tmp % git-cvsimport -v
> -d:pserver:anonymous@anoncvs.enlightenment.org:/var/cvs/e e17
> connect error: Network is unreachable
> WARNING: malformed CVS version str: Server:
> WARNING: Your CVS client version:
> [Client: Concurrent Versions System (CVS) 1.12.12 (client)]
> and/or server version:
> [Server: ]
> are too old to properly support the rlog command.
> This command was introduced in 1.11.1.  Cvsps
> will use log instead, but PatchSet numbering
> may become unstable due to pruned empty
> directories.

The cvs server is strange -- buggy probably. cvsps thinks it is old, but it is not even returning a version string. Is it really cvs?

> cvs [log aborted]: end of file from server (consult above messages if any)

The remote end is dying on you.

> I don't why it tell me that Network is unreachable as I can do normal
> cvs checkout.

I think it is the remote end that is saying Network unreachable. Perhaps it is a cvs proxy that only allows some commands?

> Also, I have the repo in another directory, and I don't know how to
> use a :local: CVSROOT

   git-cvsimport -d /tmp/cvsrepo/ modulename

should work ok. You should also try parsecvs ;-)

Pavel Roskin· May 18, 2006, 03:59 UTC · re: Martin Langhoff · lore

Re: cvsimport weird

On Thu, 2006-05-18 at 14:54 +1200, Martin Langhoff wrote:

> On 5/18/06, Bertrand Jacquin <beber.mailing@gmail.com> wrote:
> >
> The cvs server is strange -- buggy probably. cvsps thinks it is old,
> but it is not even returning a version string. Is it really cvs?

The version reporting is working for me:

$ cvs -d :pserver:anonymous@anoncvs.enlightenment.org:/var/cvs/e version
Client: Concurrent Versions System (CVS) 1.11.21 (client/server)
Server: Concurrent Versions System (CVS) 1.11.17 (client/server)

But I can reproduce the problem with git-cvsimport. git main branch, cvsps 2.1.

I'm quite sure that it's a bug in cvsps. It displays such things on x86_64, but works properly on 32-bit PowerPC.

x86_64:
$ cvsps --cvs-direct -A -u --root :pserver:anonymous@anoncvs.enlightenment.org:/var/cvs/e e17
connect error: Network is unreachable
WARNING: malformed CVS version: no data
WARNING: malformed CVS version str: (UNKNOWN CLIENT)
WARNING: Your CVS client version:
[(UNKNOWN CLIENT)]
and/or server version:
[(UNKNOWN SERVER)]

ppc: $ cvsps --cvs-direct -A -u --root :pserver:anonymous@anoncvs.enlightenment.org:/var/cvs/e e17 cvs_direct initialized to CVSROOT /var/cvs/e cvs rlog: Logging e17 cvs rlog: Logging e17/CVSROOT cvs rlog: Logging e17/apps cvs rlog: Logging e17/apps/e cvs rlog: Logging e17/apps/e/client ...

Both are cvsps 2.1 on Fedora Core 5.

-- 
Regards,
Pavel Roskin

Pavel Roskin· May 18, 2006, 05:44 UTC · re: Pavel Roskin · lore

Re: cvsimport weird

On Wed, 2006-05-17 at 23:59 -0400, Pavel Roskin wrote:

> I'm quite sure that it's a bug in cvsps.  It displays such things on
> x86_64, but works properly on 32-bit PowerPC.

Address resolution is broken in cvsps on 64-bit machines. This patch to cvsps is needed:

--- cbtcommon/tcpsocket.c +++ cbtcommon/tcpsocket.c

@@ -198,7 +198,7 @@ convert_address(long *dest, const char *
     memcpy(dest, &ip.s_addr, sizeof(ip.s_addr));
   }
 #else
-  if ( (*dest = inet_addr(addr_str)) != -1)
+  if ( (*dest = inet_addr(addr_str)) != INADDR_NONE)
   {
     /* nothing */
   }


However, it's not sufficient to fix the original problem of empty CVS
server version string.  For some reason cvsps fails to authenticate with
pserver.  The ext protocol is working.

It's interesting that both git-cvsimport and cvsps have code to
authenticate over the pserver protocol.  I think maybe git-cvsimport
shouldn't do it, or maybe it should close some sockets before running
cvsps.

-- 
Regards,
Pavel Roskin

YOSHIFUJI Hideaki / 吉藤英明· May 18, 2006, 05:56 UTC · re: Pavel Roskin · lore

Re: cvsimport weird

In article <1147931094.32050.51.camel@dv> (at Thu, 18 May 2006 01:44:54 -0400), Pavel Roskin <proski@gnu.org> says:

Show 14 quoted lines

> Address resolution is broken in cvsps on 64-bit machines.  This patch to
> cvsps is needed:
> 
> --- cbtcommon/tcpsocket.c
> +++ cbtcommon/tcpsocket.c
> @@ -198,7 +198,7 @@ convert_address(long *dest, const char *
>      memcpy(dest, &ip.s_addr, sizeof(ip.s_addr));
>    }
>  #else
> -  if ( (*dest = inet_addr(addr_str)) != -1)
> +  if ( (*dest = inet_addr(addr_str)) != INADDR_NONE)
>    {
>      /* nothing */
>    }

You need to define INADDR_NONE on some platforms; e.g. Solaris.

--yoshfuji

Bertrand Jacquin· May 18, 2006, 09:33 UTC · re: Pavel Roskin · lore

Re: cvsimport weird

On 5/18/06, Pavel Roskin <proski@gnu.org> wrote:

Show 6 quoted lines

> On Wed, 2006-05-17 at 23:59 -0400, Pavel Roskin wrote:
> > I'm quite sure that it's a bug in cvsps.  It displays such things on
> > x86_64, but works properly on 32-bit PowerPC.
>
> Address resolution is broken in cvsps on 64-bit machines.  This patch to
> cvsps is needed:

Ah Thanks, I'm running it on a x86_64.

-- 
Beber
#e.fr@freenode