Re: [PATCH/RFC v3 1/6] reflog: refactor parsing and checking

Hi,

On Sat, 17 Jan 2009, Thomas Rast wrote:

Note that this switches for_each_reflog_ent() from silently ignoring 
errors to die().

I am not convinced this is necessary, and neither that it is good.

You could return int from parse_reflog_line() just like the other 
non-dying functions.

And keep in mind that reflogs could be corrupted for all kinds of reasons; 
with your patch, reflog handling would fail gracelessly, even if a 
lot could still be salvaged (by ignoring the error with a warning).

Ciao,
Dscho