Re: how to skip branches on git svn clone/fetch when there are errors

From: Daniele Segato <daniele.bilug@gmail.com>
Date: Sep 7, 2009, 09:30 UTC
Message-ID: <9accb4400909070230n413c6ecfqef8238422dd5d3b@mail.gmail.com>
In-Reply-To: <9accb4400908310126v15b08c7fr425c9daff26012f3@mail.gmail.com>

On Mon, Aug 31, 2009 at 10:26 AM, Daniele Segato<daniele.bilug@gmail.com> wrote:>

Show 7 quoted lines

> git init
> git svn init svn://svn.mydomain.com/path/to/repo -T HEAD -b BRANCHES -t TAGS
> $ git svn fetch
> Use of uninitialized value in concatenation (.) or string at
> /usr/lib/perl5/SVN/Core.pm line 584.
> Authorization failed:  at /usr/bin/git-svn line 1415
>

more info on the error (enabled the confess instead of croak to the Core.pm library)

Use of uninitialized value in concatenation (.) or string at
/usr/lib/perl5/SVN/Core.pm line 585.
Authorization failed:  at /usr/lib/perl5/SVN/Core.pm line 654
	SVN::Error::confess_on_error('_p_svn_error_t=SCALAR(0x9492a50)')
called at /usr/lib/perl5/SVN/Ra.pm line 492
	SVN::Ra::AUTOLOAD('Git::SVN::Ra=HASH(0x945dae8)',
'alfresco-enterprise-mirror/alfresco/BRANCHES/V2.1-A/root', 7738,
'SVN::Pool=REF(0x9492bc0)') called at /usr/bin/git-svn line 3760
	Git::SVN::Ra::check_path('Git::SVN::Ra=HASH(0x945dae8)',
'alfresco-enterprise-mirror/alfresco/BRANCHES/V2.1-A/root', 7738)
called at /usr/bin/git-svn line 4045
	Git::SVN::Ra::get_dir_check('Git::SVN::Ra=HASH(0x945dae8)',
'HASH(0x92131e0)', 'HASH(0x9388050)', 7738) called at /usr/bin/git-svn
line 4062
	Git::SVN::Ra::match_globs('Git::SVN::Ra=HASH(0x945dae8)',
'HASH(0x92131e0)', 'HASH(0x9463c00)', 'ARRAY(0x90bbc00)', 7738) called
at /usr/bin/git-svn line 3985
	Git::SVN::Ra::gs_fetch_loop_common('Git::SVN::Ra=HASH(0x945dae8)',
7737, 16113, 'ARRAY(0x90bbbe0)', 'ARRAY(0x90bbc00)') called at
/usr/bin/git-svn line 1415
	Git::SVN::fetch_all('svn', 'HASH(0x9464250)') called at
/usr/bin/git-svn line 372
	main::cmd_fetch() called at /usr/bin/git-svn line 253
	eval {...} called at /usr/bin/git-svn line 251

I'll keep looking at it to see if I can figure out a way to "skip" the error myself and, eventually, provide a patch

Bye, Daniele