You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by Jeremy Kitchen <ki...@scriptkitchen.com> on 2008/10/17 00:34:13 UTC

Re: commit problem with 1.4.2 (r22196) [solved, bugworthy?]

Jeremy Kitchen wrote:
> The problem is this:
> "svn-commit.tmp" 6L, 137C written
> Sending        Mailfilterer.pm
> Transmitting file data .svn: Commit failed (details follow):
> svn: MERGE request failed on '/repo/trunk/perl/Ndn/Dreamhost/Service'
> svn: Valid UTF-8 data
> (hex: 4f 63 74 20 31 35 20 31 38 3a 33 30 3a 35 34 20 32 30 30 38 5d 20
> 70 60)
> followed by invalid UTF-8 sequence
> (hex: 9a 02 3a 20)
> svn: Your commit message was left in a temporary file:
> svn:    '/home/jeremyk/devel/ndn/perl/Ndn/Dreamhost/Service/svn-commit.tmp'

turns out this was actually due to our pre-commit hook script failing
and spitting out some strange output:

[Thu Oct 16 17:29:53 2008] @ߚ: Name "OpenSRS::OPENSRS_LOGOUT" used only
once: possible typo at @ߚ line 51.

Now, my question is:
Should this message be a little more clear, perhaps?  If it said
something like 'pre-commit hook script exited non-zero, but we could not
parse the output, here's why', it would have been a lot easier to
troubleshoot this issue :)

thoughts?

-Jeremy


Re: commit problem with 1.4.2 (r22196) [solved, bugworthy?]

Posted by Jeremy Kitchen <ki...@scriptkitchen.com>.
Daniel Shahaf wrote:
>> turns out this was actually due to our pre-commit hook script failing
>> and spitting out some strange output:
>>
>> [Thu Oct 16 17:29:53 2008] @ߚ: Name "OpenSRS::OPENSRS_LOGOUT" used only
>> once: possible typo at @ߚ line 51.
>>
>> Now, my question is:
>> Should this message be a little more clear, perhaps?
> 
> What, reading the hex'ed timestamp isn't clear enough already?  :-)

nope, it wasn't, unfortunately.  If I come across it again, it certainly
will be, however ;)

>> If it said
>> something like 'pre-commit hook script exited non-zero, but we could not
>> parse the output, here's why', it would have been a lot easier to
>> troubleshoot this issue :)
>>
> 
> Seems that it's fixed already: in 1.5.3, the code reads:
> 
>   /* If we got the stderr output okay, try to translate it into UTF-8.
>      Ensure there is something sensible in the UTF-8 string regardless. */
[snip code]

Excellent, thanks!

-Jeremy





Re: commit problem with 1.4.2 (r22196) [solved, bugworthy?]

Posted by Daniel Shahaf <d....@daniel.shahaf.name>.
Jeremy Kitchen wrote on Thu, 16 Oct 2008 at 17:34 -0700:
> Jeremy Kitchen wrote:
> > The problem is this:
> > "svn-commit.tmp" 6L, 137C written
> > Sending        Mailfilterer.pm
> > Transmitting file data .svn: Commit failed (details follow):
> > svn: MERGE request failed on '/repo/trunk/perl/Ndn/Dreamhost/Service'
> > svn: Valid UTF-8 data
> > (hex: 4f 63 74 20 31 35 20 31 38 3a 33 30 3a 35 34 20 32 30 30 38 5d 20
> > 70 60)
> > followed by invalid UTF-8 sequence
> > (hex: 9a 02 3a 20)
> > svn: Your commit message was left in a temporary file:
> > svn:    '/home/jeremyk/devel/ndn/perl/Ndn/Dreamhost/Service/svn-commit.tmp'
> 
> turns out this was actually due to our pre-commit hook script failing
> and spitting out some strange output:
> 
> [Thu Oct 16 17:29:53 2008] @ߚ: Name "OpenSRS::OPENSRS_LOGOUT" used only
> once: possible typo at @ߚ line 51.
> 
> Now, my question is:
> Should this message be a little more clear, perhaps?

What, reading the hex'ed timestamp isn't clear enough already?  :-)

> If it said
> something like 'pre-commit hook script exited non-zero, but we could not
> parse the output, here's why', it would have been a lot easier to
> troubleshoot this issue :)
> 

Seems that it's fixed already: in 1.5.3, the code reads:

  /* If we got the stderr output okay, try to translate it into UTF-8.
     Ensure there is something sensible in the UTF-8 string regardless. */
  if (!err2)
    {
      err2 = svn_utf_cstring_to_utf8(&utf8_stderr, native_stderr->data, pool);
      if (err2)
        utf8_stderr = _("[Error output could not be translated from the "
                        "native locale to UTF-8.]");
    }
  else
    {
      utf8_stderr = _("[Error output could not be read.]");
    }

Daniel

> thoughts?
> 
> -Jeremy
> 
>