[Buildroot] [PATCH v2] core/pkg-generic: only save latest package list

[John Keeping]

On Tue, 1 May 2018 14:04:16 +0200
"Yann E. MORIN" <yann.morin.1998 at free.fr> wrote:

> John, All,
> 
> On 2018-05-01 12:13 +0100, John Keeping spake thusly:
> > When rebuilding a package, simply appending the package's file list
> > to the global list means that the package list grows for every
> > rebuild, as does the time taken to check for files installed by
> > multiple packages. Furthermore, we get false positives where a file
> > is reported as being installed by multiple copies of the same
> > package.
> > 
> > With this approach we may end up with orphaned files in the target
> > filesystem if a package that has been updated and rebuilt no longer
> > installs the same set of files, but we know that only a clean build
> > will produce reliable results.  In fact it may be helpful to
> > identify these orphaned files as evidence that the build is not
> > clean.
> > 
> > Signed-off-by: John Keeping <john at metanate.com>
> > ---
> > Changes in v2:
> > - Do the sed as a separate step rather than in the same shell
> > invocation as the find (we know that $(BUILD_DIR) is an absolute
> > path); since we no longer ignore the exit status of sed, we also
> > have to ensure that the file exists before executing sed
> > 
> >  package/pkg-generic.mk | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/package/pkg-generic.mk b/package/pkg-generic.mk
> > index bd47ca1964..8ebbacc180 100644
> > --- a/package/pkg-generic.mk
> > +++ b/package/pkg-generic.mk
> > @@ -63,6 +63,8 @@ GLOBAL_INSTRUMENTATION_HOOKS += step_time
> >  # $(2): base directory to search in
> >  # $(3): suffix of file  (optional)
> >  define step_pkg_size_inner
> > +	@>$(BUILD_DIR)/packages-file-list$(3).txt  
> 
> Have you tested that? In my little experience with dealing with the
> shell, this actually truncates the file to a zero-size, and thus we
> loose the existing package->file list, even for other packages.
> 
> So, just do as I suggested: use touch to ensure the file exists.
> 
> (Yeah, one could also use the append-redirection >> instead of touch,
> but is that trick really an improvement over touch? touch is well
> known, and virtually everyone will recognise a construct that uses
> touch before sed-ing a file.)

Oops, you're right, that will truncate the file.  I'm used to the Git
guidelines of avoiding touch when we don't care about the mtime of the
file, but I think it's fine here.

> > +	$(SED) '/^$(1),/d' $(BUILD_DIR)/packages-file-list$(3).txt
> >  	cd $(2); \
> >  	find . \( -type f -o -type l \) \
> >  		-newer $($(PKG)_DIR)/.stamp_built \
> > -- 
> > 2.17.0
> >   
>