[pacman-dev] [PATCH v2 6/7] start hook documentation

[Andrew Gregory]

On 09/19/15 at 06:25pm, Allan McRae wrote:
> On 15/09/15 08:42, Andrew Gregory wrote:
> > ---
> >  doc/.gitignore       |   1 +
> >  doc/Makefile.am      |   4 ++
> >  doc/alpm-hooks.5.txt | 107 +++++++++++++++++++++++++++++++++++++++++++++++++++
> >  3 files changed, 112 insertions(+)
> >  create mode 100644 doc/alpm-hooks.5.txt
> > 
> 
> libalpm-hooks vs alpm-hooks...  I lean towards using libalpm-hooks.
> Especially as we already have libalpm(3).

<snip>

> > +CAVEATS
> > +-------
> > +
> > +There is no way to guarantee that the trigger operation was actually performed
> > +for file triggers.  Removal triggers will match even if the file did not
> > +actually exist on the file system when the transaction begin.  
> 
> began
> 
> 
> > Sync triggers
> > +may be extracted as a .pacnew file, leaving the trigger path unmodified.
> > +
> 
> Does that mean /etc/foo.pacnew will not match a Trigger /etc/foo?  Or
> will it match despite the file not being modified?

Hooks are completely unaware of backup handling so anything .pacnew
will not match.  What this is trying to say is that if you have
a trigger for '/etc/foo', pacman may ultimately extract it as
'/etc/foo.pacnew' leaving '/etc/foo' untouched but a SYNC hook for
'/etc/foo' would still be triggered.

> That section needs to be clearer.
> 
> > +include::footer.txt[]