Get ReadHMF errors to report full-path to faulty Holmakefile

Read Holmakefiles fewer times, speeding up startup times

This speeds up Holmake, hol and script files.

Also remove output of loadPath deltas as hol starts.

Make interactive HOL recursively read INCLUDES to augment loadPath

Closes #604

Ensure interactive loadPaths are adjusted with holpathdb info in scope

Thanks to Ramana Kumar for reporting the problem

Make polyscripter tool pay attention to Holmakefile INCLUDES line

Refactor some code along the way so that this reading of Holmakefiles
is now a standard facility in the ReadHMF structure.

Test that it works by making the polyscripter test read from
examples/misc.

Read .holpath directories to allow relocatable developments

If there is a .holpath file in a directory containing a string with a
name for the development under that directory, then the object
files (.uo and .ui) in that directory will be easily relocatable
because they won't use absolute paths, but will instead use
$(devname)/ type paths.

Multiple developments can combine in this way, and there is additional
machinery to make sure that all of the relevant directories are
scanned before reading and writing of object files happens.

Correct a bug in Holmake’s handling of nested conditional directives.

Problem was that if the top conditional evaluated to false, as in

ifeq(1,0)
ifdef FOO
endif
endif

the presence of the inner ifdef test wasn’t recognised as such, and so
the second endif was seen as superfluous, giving an error.

Fix a minor 'definedness' bug in Holmake's conditional directives.

This is progress with issue #11.

In particular, variables have to be considered undefined if one
expansion through the environment gives back either a fragment list
consisting of [LIT ""], or just the empty list.

Implement ifeq and ifneq conditionals for Holmake.

Still need to document these. This is progress with issue #11.

Use line-numbered error messages when failing in ReadHMF.

Make ifneq skipped when "Skipping Elses" in ReadHMF.

Implement ifdef/ifndef conditional directives for Holmake. (Progress for #11.)

For the moment, only ifdef and ifndef are supported. It'd be good to
also do ifeq and ifneq. The behaviour should be as per GNU make.

In other words, if your Holmakefile looks like

ifdef POLY
mytarget: mydeps
command1
command2
endif

then the delimited text will only be seen if the POLY variable has a
non-empty value (which will happen if you're running Poly/ML HOL, or
if you've got an environment variable called POLY).

You can also use 'else' to do things if the head condition is false.
Thus:

ifdef POLY
text...
else
alternative text
endif

Finally, these directives can be nested, and you can chain else-if on
the same line:

ifdef POLY
text...
else ifdef FOOBAR
text...
else
text...
endif

Get mosml Holmake to build with new Holmake_types.

Delete unnecessary case in pattern-match.

TextIO.inputLine also returns at least a newline when it returns SOME
of anything. (I.e., even if the file doesn't end with a newline, it
will return the last line with an extra newline tacked on.)

Treat \r\n as \n, even on Unix (helps with Poly on cygwin).

Fix startup problem for Poly/ML HOL (caused by recent Holmake refactoring).

Get Poly and Moscow ML implementations of Holmake to share more code.

There is still a lot more that could be done here.

Start to implement the Basis 97 extensions that Moscow ML hasn't got
in order to force our codebase to get up-to-date. It should also
mean less bodging around for the Poly/ML code. I haven't checked that
my changes to tools-poly/poly/poly-init2.ML have done all that is
required yet. Feel free to fix problems arising there (I hope it will
just be a matter of deleting things).

Rework lexing of Holmakefiles so that the system doesn't choke on
Joe's monster example. The result is simpler code (which is a bit
worrying in itself), and one less dependency on mosmllex. It
successfully reads all of the distribution's makefiles, as well as
Joe's.