* libguile/load.c (scm_init_load_path): If GUILE_SYSTEM_PATH is set, use
that instead of the compiled-in suffix to the load path. And, as a
special case, GUILE_SYSTEM_PATH= is interpreted as '(). A bit nasty.
* pre-inst-guile-env.in (top_builddir): Set GUILE_SYSTEM_PATH to the
empty string, if it is not set.
* libguile/vm-engine.c (vm_run): Add new error case,
vm_error_not_enough_values.
* libguile/vm-i-system.c (goto/nargs, call/nargs): So, in these cases, if
we get too many values, we don't truncate the values like we do in the
single-value continuation case, or in the mvbind case. What to do? I
guess we either truncate them here, or only allow the correct number of
values. Dunno. Mark the code as a fixme.
(truncate-values): New instruction, for mv-bind: checks that the number
of values on the stack is compatible with the number of bindings we
have arranged for them, truncating if necessary.
* module/language/scheme/translate.scm (custom-transformer-table):
Compile receive as a primary form -- not so much because it is a
primary form, but more to test the mv-bind machinery. Also it's more
efficient, I think.
* module/system/il/compile.scm (lift-variables!): New helper, factored
out of `optimize'.
(optimize): Add a few more cases. Adapt `lambda' optimization, which
isn't much. I'm not happy with ghil as a mungeable language.
Add a case for call-with-values with the second argument is
a lambda: lift the lambda. Untested.
(codegen): Refactor the push-bindings! code. Compile mv-bind.
* module/system/il/ghil.scm (<ghil-mv-bind>): Add mv-bind construct,
along with its procedures.
* module/system/il/glil.scm (<glil-mv-bind>): Add mv-bind construct,
different from the high-level one. It makes sense in the source, I
think.
* module/system/vm/assemble.scm (codegen): Assemble glil-mv-bind by
pushing onto the bindings list, and actually push some code to truncate
the values.
Hello!
The attached patch fixes first-time compilation, by ensuring SRFI
modules are built before "guile-tools compile" is ever run.
Thanks,
Ludo'.
From 691a111c440a26c021f52b4027b0d9772f8e04cc Mon Sep 17 00:00:00 2001
From: =?utf-8?q?Ludovic=20Court=C3=A8s?= <ludo@gnu.org>
Date: Tue, 16 Sep 2008 23:13:38 +0200
Subject: [PATCH] Fix compilation order of the sub-directories.
* Makefile.am (SUBDIRS): Move `ice-9' past `srfi' and friends, so that
the SRFI modules needed by the compiler are built before "guile-tools
compile" is used.
* libguile/vm-engine.c (vm_run): Add another byte onto the bootstrap
program, as the offset passed to mv-call now takes two bytes.
* module/system/vm/frame.scm (bootstrap-frame?): Update for the new
bootstrap length. Really we should just check for 'halt though.
* libguile/vm-i-system.c (FETCH_OFFSET): New helper, used in BR().
(goto/nargs, call/nargs): Versions of goto/args and call, respectively,
that take the number of arguments from a value on the top of the stack.
(mv-call): Call FETCH_OFFSET to get the offset.
* module/language/scheme/translate.scm (custom-transformer-table):
Compile call-with-values to <ghil-mv-call>. There is some trickery
because of the r4rs.scm call-with-values trampolines.
* module/system/il/ghil.scm: Add <ghil-mv-call> and accessors.
* module/system/il/compile.scm (codegen): Compile <ghil-mv-call>.
* module/system/il/glil.scm: Add <glil-mv-call>, which needs some special
assembly because of the label. Fix some typos.
* module/system/vm/assemble.scm (byte-length): New helper, factored out
and made more general.
(codegen): Assemble mv-call, including the label.
(check-length): New helper, makes sure that the addressing is
consistent within the produced object code.
(stack->bytes): Rewrite to be more generic -- now `br' instructions
aren't the only ones jumping around in the instruction stream.
* module/system/vm/conv.scm (make-byte-decoder): Return two values in the
#f case.
* module/system/vm/disasm.scm (disassemble-bytecode): Rewrite, because
the previous implementation depended on a guile interpreter quirk:
namely, that multiple values could be represented within one value, and
destructured later.
* libguile/vm-engine.c (vm_run): Move nvalues to the top level, to avoid
(spurious, it seems) gcc warnings about it being used uninitialized.
* libguile/vm-i-system.c (halt, return/values): Adapt to gcc silliness.
Deindent some of return/values.
(return/values*): New instruction, does what (apply values . args)
would do.
* module/language/scheme/translate.scm (custom-transformer-table): Move
the apply and @apply cases here from inline.scm, because we need some
more cleverness when dealing with cases like (apply values . args).
(lookup-apply-transformer): Define an eval transformer for `values',
turning it into ghil-values*.
* module/system/il/compile.scm (codegen): Compile <ghil-values*> into
return/values*.
* module/system/il/ghil.scm: Add <ghil-values*> and accessors.
(ghil-lookup): Add optional argument, define?, which if false tells us
not to actually cache the binding if it is not found in the toplevel.
* module/system/il/inline.scm: Remove apply clauses.
* module/system/vm/frame.scm (bootstrap-frame?): Update heuristic for
bootstrap-frame?, as the bootstrap frame is now 5 bytes since it
accepts multiple values.
* libguile/vm-engine.c (vm_run): The bootstrap program now uses mv_call,
so as to allow multiple values out of the VM. (It did before, because
multiple values were represented internally as single scm_values
objects, but now that values go on the stack, we need to note the boot
frame as accepting multiple values.)
(vm_error_no_values): New error, happens if you pass no values into a
single-value continuation. Passing more than one is OK though, it just
takes the first one.
* libguile/vm-i-system.c (halt): Assume that someone has pushed the
number of values onto the stack, and package up that number of values
as a scm_values() object, for communication with the interpreter.
(mv-call): New instruction, calls a procedure with a multiple-value
continuation, even handling calls out to the interpreter.
(return/values): New instruction, returns multiple values to the
continuation. If the continuation is single-valued, takes the first
value or errors if there are no values. Otherwise it returns to the
multiple-value return address, pushing the number of values on top of
the values.
* module/system/il/compile.scm (codegen): Compile <ghil-values> forms.
* module/system/il/ghil.scm (<ghil-values>) Add new GHIL data structure
and associated procedures.
* module/language/scheme/translate.scm (custom-transformer-table):
Compile (values .. ) forms into <ghil-values>.
* module/language/scheme/translate.scm (custom-transformer-table): Rename
from `primitive-syntax-table', because now it will handle procedural
values as well.
(lookup-transformer): Update for renaming. Look up custom transformers
by value, not name.
(make-pmatch-transformers): Key the transformer table by value, not
name.
* libguile/vm-i-system.c (call): Rename continuation invocation from
`vm_call_cc' to `vm_call_continuation', because that's what it really
does. Add a note that it doesn't handle multiple values at the moment.
(goto/arg): Renamed from tail-call, in deference to the progenitors, on
Dale Smith's suggestion.
(goto/apply): New instruction, for `apply' in a tail context. Not yet
used, or vetted for that matter.
(call/cc): No need to pop the program, I don't think; although this
isn't tested either.
(goto/cc): New instruction, for call/cc in a tail context.
* module/language/scheme/translate.scm (*forbidden-primitives*): Rename
from %forbidden-primitives.
* module/system/il/compile.scm (codegen): Adapt to goto/args instead of
tail-call.
* module/system/il/inline.scm: Start inlining some macros used in
r4rs.scm -- not yet fully tested.
* ice-9/boot-9.scm: Allow load of a compiled r4rs file.
* libguile/frames.c (frame-mv-return-address): New accessor.
* libguile/frames.h: Update frame diagram.
(SCM_FRAME_UPPER_ADDRESS): Update for data area
growing by one pointer.
(SCM_FRAME_MV_RETURN_ADDRESS): New macro.
* libguile/vm-engine.h (NEW_FRAME): Update for frame getting bigger by a
pointer. In a normal NEW_FRAME, set the MV return address to NULL, to
indicate that this continuation does not accept multiple values.
* libguile/vm-i-system.c (tail-call): Update frame replacement code to
understand the MV return address.
(return): Make room for the MVRA.
* libguile/programs.c (program_print): Only try to lookup write-program
if the module system is booted.
* libguile/vm-engine.h (FREE_FRAME): Remove, it's now inlined everywhere.
* libguile/vm-i-system.c (tail-call): Inline FREE_FRAME, and implement
the calling bits here. Will make things more hackable.
* module/system/vm/program.scm (program-bindings-as-lambda-list): Handle
the bindings-is-null case too -- not sure how it comes about, though. A
thunk with no let, perhaps.
(write-program): Another default for the name: the source location at
which it was defined.
* libguile/programs.c (program_print): Add some "logic" to stop doing
detailed prints if one print had a nonlocal exit -- preventing
exceptions in backtraces.
* module/system/vm/frame.scm (frame-call-representation)
(frame-program-name): Rename program-name to frame-program-name, and
use the program-name if it is available.
* module/system/vm/program.scm (program-bindings): Return #f if there are
no bindings.
(program-name): New public procedure.
(program-bindings-as-lambda-list, write-program): A more useful writer
for programs.
* libguile/programs.c (scm_bootstrap_programs, program_print): Add a smob
printer for programs, which dispatches to `write-program'.
* module/language/scheme/translate.scm (primitive-syntax-table): In forms
like (define x y) where y is a lambda, and the lambda has no name yet,
set the lambda's name in its metadata.
* module/system/il/compile.scm (codegen): Fix *another* bug in compiling
`or' -- in the case in which the value was being discarded, as in `or'
used as a control structure, we were sometimes leaving a value on the
stack.
* testsuite/t-or.scm: Add another test case for `or'.
* ice-9/boot-9.scm (@, @@): Note that these don't work with the compiler.
Damn.
(top-repl): Run the VM repl. Whooo!
* module/system/repl/repl.scm (start-repl): Catch 'quit, as the
scm-style-repl does. Newline after input EOF's, so that we don't leave
the user's shell messed up.
* module/language/scheme/spec.scm (scheme): Specify an evaluator, `eval'.
* module/system/repl/common.scm (repl-default-options): Add option,
`interp', specifying that, if possible, the repl should interpret its
expressions rather than compile them. Defaults to #f.
* libguile/programs.h (struct scm_program):
* libguile/programs.c (scm_c_make_program): Record the current module
when making a program. This replaces the per-late binding recorded
module in the generated code, which should be more efficient, both in
terms of garbage, and in not calling resolve-module.
(program-module): New accessor.
* module/system/vm/program.scm: Add program-module to exports.
* libguile/vm-i-loader.c (link-later): Remove this instruction, since now
the entry in the object table is just a symbol, and can be loaded with
load-symbol.
* libguile/vm-i-system.c (late-variable-ref, late-variable-set): Rework
so as to look up in the module of the current program. The logic could
be condensed quite a bit if scm_module_lookup () knew what to do with
mod==#f.
* module/system/vm/assemble.scm (dump-object!): Dump <vlink-later> just
as load-symbol, as mentioned in the note on link-later.
* module/system/il/ghil.scm: Update comment to reflect the new reality.
* module/system/base/syntax.scm (keywords): Don't enable :keywords, it
breaks code that may assume that ':foo is a symbol, like boot-9.
* module/*.scm: Don't use :keywords, use #:keywords. The user can decide
if she wants #:keywords in their .guile, and :keywords might make us
compile modules differently.
* module/system/vm/assemble.scm (make-meta, codegen): Hide the "meta"
information -- the names of the bindings, source info, procedure
properties, etc -- behind a lambda. This way, loading up a program
conses less, because the metadata stays as mmap'd code until it is
needed.
* libguile/vm-i-loader.c (load-program): Adjust load-program to expect
the metadata to be a program.
* module/system/vm/program.scm (program-bindings, program-sources)
(program-properties): Adjust to new meta format.
* ice-9/Makefile.am (SOURCES, NOCOMP_SOURCES): Compile boot-9.scm.
Wooooo! This makes some things harder to debug, and program loading
needs to cons much less, but I think it makes sense to compile boot-9
by default if for no other reason than to catch bugs earlier.
* ice-9/boot-9.scm (defmacro:syntax-transformer, define-syntax-macro):
Removed these, as I could not see anywhere they were being used, and
they use the unnecessary procedure->syntax procedure.
* module/language/scheme/translate.scm (%forbidden-primitives): Take
procedure->memoizing-macro off probation; although it's not a good
idea, there is a fair amount of existing code that uses it that can be
compiled fine. So allow it in that case.
* ice-9/boot-9.scm: Postpone module system boot until (%app modules) is
defined, so that resolve-module will work. This might not actually be
necessary given the previous tomfoolery in resolve-module, but it
doesn't seem like a bad change.
* libguile/modules.c (module-variable): If module is #f, access the
pre-modules-array. This is so that nested-ref can work before the
module system is booted, I think.
Of course all of these dependency lines during bootstrap are just to
make sure the system can be booted properly, either interpreted or
compiled, so there's no one right way: there are many ways that could
work.
* ice-9/boot-9.scm (resolve-module): Change so that resolving '(guile)
does not require any module lookups. This is so that while within a
call to (resolve-module '(guile)), we don't recurse when looking up the
location for e.g. `append'. I can imagine other ways to get around
this, but this one seems OK.
* libguile/vm-i-system.c (late-variable-ref, late-variable-set): If the
module system isn't booted, do a simple scm_lookup. In the -ref case,
actually cache the variable location (doh!).
* module/system/il/compile.scm (constant?, codegen): Add some diagnostics
so that we can get decent error reporting if we accidentally unquote an
unreadable value into the compiled output.
* module/system/il/ghil.scm (<ghil-env>, <ghil-toplevel-env>): Refactor
so that all environments point (eventually) at one toplevel
environment. Instead of having potentially multiple toplevel
environments, each noting the module against which its bindings are
resolved, have each binding in the toplevel record what module it
should be resolved in. Should fix compilation units that define
multiple modules.
(ghil-lookup, ghil-define): Reworked to not be destructive. Module
variables now have the module name as their "env", and are keyed as
`(MODNAME . SYM)' in the var table.
(call-with-ghil-environment): Reindented.
* module/system/il/inline.scm (try-inline-with-env): Adapt to
env/toplevel changes.
* module/system/vm/assemble.scm (dump-object!): A vlink-later now holds
the module name, not the module itself.
* module/system/il/compile.scm (make-glil-var): The "env" of a "module"
var is now the module name, not the module.
* module/language/scheme/translate.scm (primitive-syntax-table): Update
the way we test for toplevel environments. Reindent the lambda
translator.
(lookup-transformer, trans): lookup-transformer now has 2 args, not 3.
(translate): Update the way we make toplevel environments.
* module/system/vm/frame.scm (frame-call-representation): Show more of
lists.
(program-name): Avoid a traceback if (frame-address link) is #f. Not
sure when this can happen, but it does, and since this is already in
the backtrace function, there be badness there.
* libguile/load.h: Update scm_search_path prototype.
* libguile/load.c: Include vm.h for load-compiled/vm. Not sure if this is
bad wrt modularity.
(scm_c_string_has_an_ext): New private helper.
(scm_search_path): Add an extra optional arg, `require_exts'; if true,
require that the returned file name have one of the given extensions.
Changes the C API, but not the scheme API.
(scm_sys_search_load_path): Adapt to scm_search_path API change.
(primitive-load-path): Here is the craziness: load a compiled file if
found and newer than the corresponding (or not) source file.
(scm_init_load): Define %load-compiled-extensions as the list of
extensions denoting compiled files; defaults to '(".go").
* libguile/vm.c: Move 'vm-run, 'vm-error, 'debug sym initialization up to
the bootstrap phase, so they are ready if load-compiled/vm is called
before (system vm vm) is loaded.
* module/language/scheme/translate.scm (primitive-syntax-table): Disable
semantics of start-stack in compiled code. I think start-stack
semantics aren't bad, but they don't have vm-based implementations at
this point.