This is the command makeppreplay that can be run in the OnWorks free hosting provider using one of our multiple free online workstations such as Ubuntu Online, Fedora Online, Windows online emulator or MAC OS online emulator
makeppreplay -- Fast repeater of selected makepp actions
?: -?, A: -A,
--arguments-file, C: -C,
-c, D: --directory,
--dry-run, H: -h,
--help, I: -I,
--include-dir, J: --just-print, K: -k,
--keep-going, L: --log,
--log-file, M: -M,
--module, N: -n,
--no-warn, P: --profile, R: --recon,
--root-directory, S: -s,
--sed, T: -t,
--temporary, V: -V,
makeppreplay [ option ... ] [ VAR=value ] [ target ... ]
mppr [ option ... ] [ VAR=value ] [ target ... ]
Makeppreplay tries to do the same things makepp did. It does this solely based on the
remembered build info, saving much time by not reading makefiles, not parsing commands,
not scanning files and not looking at every repository file, to see whether it fits. But
unlike makepp it reuses all repository links it finds.
It does not know about phony or default targets, so you have to give it real files or
directories, as a shortcut for all files makepp has built therein. By default it builds
all files makepp built in the current or chosen directory.
This approach is only correct when there are no structural changes, i.e. no different
modules to be linked, no moved files, no different "#include" statements, which should
possibly have caused those files to be (re)built. After updating from version control, or
when building against a repository, where others may have caused such changes, you should
call makepp once after potential changes (e.g. in the morning after an overnight
repository update), to be sure your build infos are up to date.
Not reading makefiles means, it will not know any subs (Perl functions) defined in them.
For this you should put all your functions into an external module. It will also not know
about any variable settings, which is only a problem if your Perl code accesses them
(because in rule actions they have already been expanded).
This is still experimental and imperfect.
· Build up a complete dependency tree starting from the given targets and decide which
of them to build based on their build check methods. That requires the method to be
remembered in the build info.
· After implementing the previous bullet item, accept more makepp opts, like "-j" or
· Allow build cache with implicit "--no-populate-bc" (because makeppreplay doesn't
guarantee a correct build). Makepp might later publish to the build cache anything it
· Have makepp remember build info for all failed commands and for those it didn't
attempt because their dependencies failed. This is only partially possible, because
dependencies may only be found out after building the files that require them. What
should happen with (a has no build info):
Options include some of the makepp options, plus two new ones, for loading modules that
would normally be loaded from a makefile.
Read the file and parse it as possibly quoted whitespace- and/or newline-separated
Cd to the given directory before loading the makefile and trying to build the targets.
This is more or less equivalent to specifying a directory with "-F", except that
subsequent "-C", "-f", "-F", "-I" and "-R" options are interpreted relative to the new
directory, rather than the old one.
Cd up to the directory containing a RootMakeppfile.
Add directory to Perl load path @INC.
Print out a brief summary of the options.
Build as many files as possible, even if some of them have errors. By default, makepp
stops when it encounters the first error, even if there are other files that need to
be built that don't depend on the erroneous file.
Changes the name of the log file to the indicated name. By default, the log file is
called .makepp/log. This file is readable with makepplog, mppl.
Load module and import any functions it exports.
Print out commands without actually executing them.
Don't bother writing a detailed description of what was done to the log file.
Turn off the entering or leaving directory messages.
Output raw timestamps before and after each action.
Run the perl expression against the command before running it, like for preprocessor
"-s 's/ -c / -E /; s/\.o\b/.i/g'". This would transform a compilation command ("-c")
into a preprocessor ("-E") invocation, changing the .o suffix to .i, which is a
conventional suffix for preprocessed C. Or you might want to add some debug output to
your linker. If you have only one command in the rule, you can do this by
substituting it into the first space: "-s 's/ / -Wl,-t /'".
When the command is altered, nothing is remembered about this, i.e. makepp will think
you have manually modified the file, causing a rebuild.
Note that (unlike makepp) makeppreplay does not parse the command, so it won't notice
that the target filename was changed. So you must still request the .o target, even
though a .i will be built. And makeppclean won't know to remove it.
Makeppreplay modifies the build info of all files it touched and of all dependencies
it found modified. It marks them in such a way that makepp nevertheless knows that it
needs to rescan these files. But with this option you can prevent these
modifications, so that makepp will simply consider all files makeppreplay created as
Print out the version number.
Don't print any warning messages.
Probably the most common use is relinking either a program or a library after editing a
makeppreplay subdir/foo.o bin/bar
mppr subdir/foo.o lib/libbar.so
Makeppreplay looks at the following environment variable:
Any options you may want to always pass.
Use makeppreplay online using onworks.net services