Re: [PATCH] hg ignore libxl lex/yacc detritus

[Keir Fraser <keir.fraser@eu.citrix.com>]

On 21/07/2010 17:42, "Dan Magenheimer" <dan.magenheimer@oracle.com> wrote:

>> They should be regenerated if the corresponding input files
>> (libxlu_cfg.l and libxlu_cfg.y) have changed.  I think make will avoid
>> rebuilding them if they haven't.  So the make lines shouldn't be
>> commented out.
> 
> Hmmm... that means they get regenerated when an "hg update"
> is done that updates the input files, resulting in the
> files being found by hg diff, which appears to be
> what prompted my .hgignore patch.

Yes, tbh I'm not sure we can really rely on the mtimes of the input files
and the checked-in generated output files being correctly ordered in a fresh
checkout of our repository. I had to comment out the rules to generate our
virtual ACPI tables in tools/firmware/hvmloader/acpi when I used to have the
output of the ACPI compiler checked in (I've got rid of that now).


So, really, it may be we either need to remove the checked-in generated
files, or comment out the rules that build those generated files. Because we
can't rely on the file modification times being 'right' in a fresh checkout.

 -- Keir