Note: when CHECK is set, targets.log only.
targets.log will be moved to $(BUILDDIR)/reports/

Fixes access for others to targets.log

We cannot get IMAGE_OUTPATH from the build.log, and there is currently
no other mechanism. Creating a CHECK directory was a bad idea. It should
have been created only for CHECK, but it was always created when DEBUG
was not enabled. So it's better to just issue a warning.

xargs -r is a good habit usually.

Fixes: c7717836

Useful for checking changes.

If the CHECK option is enabled, then now the report directory is
generated successfully and is called "$(IMAGE_NAME).$(IMAGE_TYPE)".

SVGZ is more than a hundred times smaller than PNG.
a

Wrote that it's a bashism myself and neglected
to specify the proper shell, sigh.

This SHELL manipulation could ruin tracing
piggybacked onto it within lib/report.mk
but it's too late at this stage anyways.

The first half of both pipes was clearly a copy-pasted
initial logfile processing; the file can be of considerable
size (e.g. several megabytes) so it might be slightly more
efficient and cool (but a bit more arcane) to use bash(1)'s
process substitution along with good ol' tee(1) like this:

$ echo -e '1\n2\n3' |tee /dev/stderr 2> >(grep 2 >STDERR) |grep 1 >STDOUT
$ head STD*
==> STDERR <==
2

==> STDOUT <==
1

That "clever hack" with two rev(1) invocations
seems pretty stupid to me: too much processes
and processing instead of simple prefix removal.

One can also have that as ${var##*/} when dealing
with a shell variable and avoiding extra forks, BTW.

..making it fit 80 columns while at that
to avoid horizontal spaghetti code.

Previously, when rebuilding, the new reports were copied as a
subdirectory.

Make it uniform through a function -- with the initially desired effect
of no extra lines while building multiple images with REPORT=1
(those were useless anyways printing repeatedly the same path
that's relative to a particular builddir).

Sometimes REPORT=1 just wasn't passed to make, and the build.log
saved is noisy and harder to compare through diff(1); let's move
the implementation out of the makefile and into a standalone
script so it can be reused like this:

  BUILDDIR=$(sed -rn "s/^mki-image-scripts: export GLOBAL_BUILDDIR='(.*)'$/\1/p" build/build.log
  bin/cleanlog < build/build.log > build/reports/clean.log

See "export GLOBAL_BUILDDIR=" line in the particular log
(grabbing that one from stdin is not exactly trivial though).

"Provide iso contents file" has been a long-standing
feature request regarding both starterkits and distros
proper; release managers used to compensate this with
their own scripts.

Suggested-by: sem@, legion@ and someone else too

graphviz is usually around so this buglet has lived
a really long life: "a " "b" instead of "a" "b".

This is a mild generalization of what has been done by hand
to figure out a problem with make-initrd-propagator-0.10-alt1:
stripping anything intrinsically volatile off the build.log.

The filter set isn't perfect, and the proper logging will
involve mkimage tweaks as well (e.g., one generally isn't
interested in instrumental chroots' population that much,
so it should only be verbose at the higher debug levels).

Allow to put the log files and reports in a custom directory instead
of IMIGEDIR.

It's hard to tell a successful build from a failed one
if downstream hides the exit code; it's useless to continue
a `for' loop if a pipe shoves that to a subshell; well it seems
that a bashism is worth a thousand quirks with extra fds here.
Minor regexp enhancements are also due.

reports.mk made a bit more resilient/prudent either.

No need for .reports/ in toplevel directory even after build failure,
and a few more hasty hacking fixups are also due, shame on me.

Following m-p-d, a more involved default output directory
structure is feasible now:

  ~/out/name/date/name-date-arch.type

instead of plain

  ~/out/name-date-arch.type

This particular behaviour can be achieved by passing
SORTDIR='$(IMAGE_NAME)/$(DATE)'; note the single quotes.

Reports are also saved in this resulting structure
albeit the place is still highly debatable.

REPORT_PATH comes from a makefile;
let's use corresponding notation.

Toplevel Makefile is a bit optimistic regarding
calling reports.mk...

This one extracts the exact time-ordered list of script hooks
run while building the image.

Some commons refactoring is also due.

If there's no dot(1) then its input file should just be stored;
thanks glebfm@ for reminding about this.

bin/ prefix tried to become a permatemp(tm).  No way.

The purpose is being able to examine particular target interdependency
graph for a given image having been configured to avoid convoluted
dependencies (loops in particular).

The implementation is based on SHELL hook hint by John Graham-Cumming:
http://cmcrossroads.com/ask-mr-make/6535-tracing-rule-execution-in-gnu-make