| Commit message (Expand) | Author | Age | Files | Lines |
|---|
| * | Add allow-legacy-syntax-objects? parameter...* module/ice-9/psyntax.scm (syntax?): Only recognize legacy syntax
objects if the new allow-legacy-syntax-objects? parameter is true.
* module/ice-9/boot-9.scm (allow-legacy-syntax-objects?): New
parameter.
* doc/ref/api-macros.texi (Syntax Transformer Helpers): Document the
horrible situation with legacy syntax objects.
* NEWS: Add entry.
| Andy Wingo | 2017-03-28 | 2 | -2/+10 |
| * | Psyntax generates new syntax objects...* module/ice-9/psyntax.scm (make-syntax-object): Change to make
new-style syntax objects.
* module/ice-9/psyntax-pp.scm: Regenerate.
* module/ice-9/compile-psyntax.scm (squeeze-syntax-object): Change to be
functional.
(squeeze-constant): Likewise.
(squeeze-tree-il): Likewise.
(translate-literal-syntax-objects): New pass. The compiler can embed
literal syntax objects into compiled objects, but syntax can no longer
be read/written; otherwise users could forge syntax objects. So for
the bootstrap phase, rewrite literal constants to calls to
make-syntax.
| Andy Wingo | 2017-03-28 | 3 | -848/+941 |
| * | Beginnings of psyntax switch to new syntax objects...* module/ice-9/psyntax.scm: Baby steps towards support of a new
representation of syntax objects.
* module/ice-9/psyntax-pp.scm: Regenerate.
| Andy Wingo | 2017-03-28 | 2 | -2613/+2642 |
| * | Add disjoint syntax object type...* libguile/Makefile.am (libguile_@GUILE_EFFECTIVE_VERSION@_la_SOURCES):
(DOT_X_FILES, DOT_DOC_FILES, noinst_HEADERS): Add syntax.c and
syntax.h.
* libguile/evalext.c (scm_self_evaluating_p):
* libguile/goops.c (class_syntax, scm_class_of, scm_goops_early_init):
* libguile/init.c (scm_init_guile):
* libguile/print.c (iprin1):
* libguile/tags.h (scm_tc7_syntax):
* module/oop/goops.scm (<syntax>):
* module/system/base/types.scm (%tc7-syntax, cell->object):
* module/system/vm/disassembler.scm (code-annotation): Wire up the new
data type.
* libguile/syntax.c:
* libguile/syntax.h: New files.
* module/ice-9/boot-9.scm: Move new definitions to (system syntax
internal).
* module/system/syntax.scm (print-syntax): New helper.
* module/system/vm/assembler.scm (statically-allocatable?)
(intern-constant, link-data): Arrange to be able to write syntax
objects into images.
* module/language/cps/types.scm (&syntax): New type. Remove
&hash-table; it was never detected, an internal binding, and we need
the bit to avoid going into bignum territory.
| Andy Wingo | 2017-03-28 | 7 | -7/+51 |
| * | Plumbing changes to rename "syntax-module"...* module/ice-9/psyntax.scm (%syntax-module): Rename from syntax-module
in order to make room for a new syntax-module primitive binding.
* module/ice-9/psyntax-pp.scm: Regenerate.
* module/ice-9/boot-9.scm: Push syntax bindings into an internal
module.
* module/system/syntax.scm: New file.
* module/Makefile.am (SOURCES): Add system/syntax.scm.
| Andy Wingo | 2017-03-28 | 5 | -7/+35 |
| * | i18n: add international sign positions to %locale-dump...* module/ice-9/i18n.scm (%locale-dump): modified
| Mike Gran | 2017-03-20 | 1 | -2/+6 |
| * | i18n: add debugging helper procedure for locales...* module/ice-9/i18n.scm (%locale-dump): new procedure
| Mike Gran | 2017-03-20 | 1 | -1/+67 |
| * | i18n: rename locale-monetary-digit-grouping to locale-monetary-grouping...* module/ice-9/i18n.scm (locale-monetary-digit-grouping): renamed to locale-monetary grouping
(monetary-amount->locale-string): use renamed procedure
* test-suite/tests/i18n.test (%french-locale): add LC_MONETARY
(%french-utf8-locale): add LC_MONETARY
("nl-langinfo et al."): tests for locale-monetary-grouping
| Mike Gran | 2017-03-20 | 1 | -2/+2 |
| * | web: Avoid deprecated '_IOFBF'....* module/web/client.scm (open-socket-for-uri): Use 'block instead of
_IOFBF.
| Ludovic Courtès | 2017-03-17 | 1 | -1/+1 |
| * | web: Remove export of nonexistent 'open-connection-for-uri'....* module/web/client.scm: Don't export 'open-connection-for-uri', which
doesn't exist.
| Ludovic Courtès | 2017-03-17 | 1 | -2/+1 |
| * | Update --version and REPL copyright years...* module/ice-9/command-line.scm (version-etc):
* module/system/repl/common.scm (*version*): Update release year.
| Andy Wingo | 2017-03-15 | 2 | -3/+3 |
| * | Nonlocal prompt returns cause all effects...* module/language/cps/effects-analysis.scm (expression-effects): Prompts
cause &all-effects. I tried to limit this change to CSE but it was
actually LICM that was borked, so better to be conservative
* test-suite/tests/control.test ("escape-only continuations"): Add
test.
| Andy Wingo | 2017-03-13 | 1 | -1/+4 |
| * | VM support for string-set!; slimmer read-string...* doc/ref/vm.texi (Inlined Scheme Instructions): Add string-set!.
* libguile/vm-engine.c (string-set!): New opcode.
* module/ice-9/rdelim.scm (read-string): Reimplement in terms of a
geometrically growing list of strings, to reduce total heap usage when
reading big files.
* module/language/cps/compile-bytecode.scm (compile-function): Add
string-set! support.
* module/language/cps/types.scm (string-set!): Update for &u64 index.
* module/language/tree-il/compile-cps.scm (convert): Unbox index to
string-set!.
* module/system/vm/assembler.scm (system): Export string-set!.
| Andy Wingo | 2017-03-09 | 5 | -9/+20 |
| * | Fix bug in comparison between real and complex...This bug was introduced by 35a90592501ebde7e7ddbf2486ca9d315e317d09.
* module/language/cps/specialize-numbers.scm (specialize-operations):
Check that both operands are real as a condition for
specialize-f64-comparison.
* test-suite/tests/numbers.test: Add test.
| Daniel Llorens | 2017-03-09 | 1 | -6/+8 |
| * | Remove contification restriction in case-lambda...* module/language/cps/compile-bytecode.scm (compile-function): Check for
fallthrough after $kclause too; possible to need to jump if clause
tails are contified.
* module/language/cps/contification.scm (compute-contification-candidates):
Enable inter-clause contification.
| Andy Wingo | 2017-03-09 | 2 | -31/+12 |
| * | All clauses of function have same nlocals...* module/language/cps/compile-bytecode.scm (compile-function):
* module/language/cps/slot-allocation.scm ($allocation)
(lookup-nlocals, compute-frame-size, allocate-slots): Adapt to
have one frame size per function, for all clauses.
| Andy Wingo | 2017-03-09 | 2 | -33/+25 |
| * | Adapt gcprof to preemptive interrupts...* module/statprof.scm (gcprof): Remove handle-interrupts trampoline from
captured stacks.
| Andy Wingo | 2017-03-09 | 1 | -4/+4 |
| * | Adapt statprof to new preemptive interrupts...* module/statprof.scm (profile-signal-handler): Cut an additional stack
frame, corresponding to the handle-interrupts trampoline added
recently.
| Andy Wingo | 2017-03-09 | 1 | -2/+4 |
| * | statprof: 'with-statprof' honors #:display-style....* module/statprof.scm (with-statprof): Pass #:display-style to 'statprof'.
| Ludovic Courtès | 2017-03-09 | 1 | -1/+5 |
| * | 64KB segment alignment...* module/system/vm/linker.scm (*lcm-page-size*): Rename
from *page-size*, change to 64 KB.
* libguile/loader.c (load_thunk_from_memory): Only require page size
alignment, knowing that although Guile might emit ELF with 64k
alignment, it only really needs page alignment.
| Andy Wingo | 2017-03-08 | 1 | -2/+37 |
| * | psyntax: Generate identifiers in a deterministic fashion....Fixes <http://bugs.gnu.org/20272>.
* module/ice-9/boot-9.scm (module-generate-unique-id!)
(module-gensym): New procedures.
(module): Add 'next-unique-id' field.
(the-root-module): Inherit 'next-unique-id' value from early stub.
(make-module, make-autoload-interface): Adjust calls to
module-constructor.
* module/ice-9/psyntax.scm (gen-label, new-mark): Generate unique
identifiers from the module name and the per-module unique-id.
(build-lexical-var, generate-temporaries): Use
'module-gensym' instead of 'gensym'.
* module/ice-9/psyntax-pp.scm: Regenerate.
* module/language/tree-il/fix-letrec.scm (fix-letrec!): Use
'module-gensym' instead of 'gensym'.
* module/system/base/syntax.scm (define-record): Likewise.
(transform-record): Likewise.
Co-authored-by: Ludovic Courtès <ludo@gnu.org>
| Mark H Weaver | 2017-03-07 | 5 | -50/+143 |
| * | i18n: 'number->locale-string' guesses the minimum number of decimals....This feature was removed by 4aead68cdb86ca60cc372f0cd558cadda90ddec5.
* module/ice-9/i18n.scm (number-decimal-string): Rewrite the case where
DIGIT-COUNT is not an integer.
(number->locale-string): Always pass FRACTION-DIGITS to
'number-decimal-string'.
* test-suite/tests/format.test ("~h localized number")["1234.5"]
["padding", "padchar"]: Remove decimal specifier.
* test-suite/tests/i18n.test ("number->locale-string")
["fraction",
* test-suite/tests/i18n.test ("format ~h")["12 345,678"]: Remove decimal
specifier. Remove one decimal.
* doc/ref/api-i18n.texi (Number Input and Output): Update
'number->locale-string' doc to mention the number of decimals.
| Ludovic Courtès | 2017-03-01 | 1 | -13/+27 |
| * | i18n: Fix corner cases for monetary and number string conversions....Fixes <http://bugs.gnu.org/24990>.
Reported by Martin Michel <dev@famic.de>.
* module/ice-9/i18n.scm (integer->string, number-decimal-string): New
procedures.
(monetary-amount->locale-string): Use them instead of 'number->string'
followed by 'string-split'.
(number->locale-string): Likewise.
* test-suite/tests/i18n.test ("number->locale-string")["fraction"]: Add
second argument to 'number->locale-string'.
["fraction, 1 digit"]: Round up.
["fraction, 10 digits", "trailing zeros", "negative integer"]: New
tests.
* test-suite/tests/i18n.test ("format ~h"): Pass the number of decimals
for ~h.
("monetary-amount->locale-string")["French"]: Always expect two decimals
after the comma.
["one cent", "very little money"]: New tests.
* test-suite/tests/format.test ("~h localized number")["1234.5"]:
Specify the number of decimals explicitly.
["padding"]: Expect zero decimals.
["padchar"]: Ask for one decimal.
["decimals", "locale"]: Adjust rounding.
| Ludovic Courtès | 2017-03-01 | 1 | -14/+43 |
| * | i18n: Do not represent zero as "-0"....Partly fixes <http://bugs.gnu.org/24990>.
Reported by Martin Michel <dev@famic.de>.
* module/ice-9/i18n.scm (monetary-amount->locale-string): Don't negate
AMOUNT when it's zero.
(number->locale-string): Likewise.
* test-suite/tests/i18n.test ("number->locale-string")["positive inexact
zero, 1 digit"]: New test.
("monetary-amount->locale-string")["positive inexact zero"]: New test.
| Ludovic Courtès | 2017-03-01 | 1 | -3/+4 |
| * | Recognize sh3 as compilation targets...* module/system/base/target.scm (cpu-endianness, triplet-pointer-size):
Add case for "sh3".
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
| John Paul Adrian Glaubitz | 2017-03-01 | 1 | -2/+2 |
| * | scandir: Avoid 'stat' calls on each entry....* module/ice-9/ftw.scm (scandir): Rewrite in terms of 'readdir'.
| Ludovic Courtès | 2017-03-01 | 1 | -32/+26 |
| * | REPL Server: Guard against HTTP inter-protocol exploitation attacks....Reported by Christopher Allan Webber <cwebber@dustycloud.org>
Co-authored-by: Ludovic Courtès <ludo@gnu.org>
This commit adds protection to Guile's REPL servers against HTTP
inter-protocol exploitation attacks, a scenario whereby an attacker can,
via an HTML page, cause a web browser to send data to TCP servers
listening on a loopback interface or private network. See
<https://en.wikipedia.org/wiki/Inter-protocol_exploitation> and
<https://www.jochentopf.com/hfpa/hfpa.pdf>, The HTML Form Protocol
Attack (2001) by Tochen Topf <jochen@remote.org>.
Here we add a procedure to 'before-read-hook' that looks for a possible
HTTP request-line in the first line of input from the client socket. If
present, the socket is drained and closed, and a loud warning is written
to stderr (POSIX file descriptor 2).
* module/system/repl/server.scm: Add 'maybe-check-for-http-request'
to 'before-read-hook' when this module is loaded.
(with-temporary-port-encoding, with-saved-port-line+column)
(drain-input-and-close, permissive-http-request-line?)
(check-for-http-request, guard-against-http-request)
(maybe-check-for-http-request): New procedures.
(serve-client): Use 'guard-against-http-request'.
* module/system/repl/coop-server.scm (start-repl-client): Use
'guard-against-http-request'.
* doc/ref/guile-invoke.texi (Command-line Options): In the description
of the --listen option, make the security warning more prominent.
Mention the new protection added here. Recommend using UNIX domain
sockets for REPL servers. "a path to" => "the file name of".
| Mark H Weaver | 2017-03-01 | 2 | -4/+188 |
| * | http: Do not use 'eq?' to compare characters in parse-request-uri....* module/web/http.scm (parse-request-uri): Use 'eqv?' to compare
characters.
| Mark H Weaver | 2017-03-01 | 1 | -1/+1 |
| * | texinfo: Remove unnecessary (oop goops) dependency....* module/texinfo/string-utils.scm: Remove #:use-module (oop goops).
| Ludovic Courtès | 2017-03-01 | 1 | -1/+0 |
| * | Cosmetic REPL server improvements...* module/system/repl/server.scm (run-server*): Cosmetic improvements.
| Andy Wingo | 2017-03-01 | 1 | -10/+10 |
| * | Fix trap frame matching...* module/system/vm/traps.scm (program-last-ip): Actually return an
absolute IP. Fixes traps!
| Andy Wingo | 2017-03-01 | 1 | -1/+3 |
| * | Fix class slot allocation since GOOPS rewrite...* module/oop/goops.scm (%compute-layout): Fix class slot layout.
Before, a #:class that was an argument to #:allocation was getting
interpreted as a keyword with a value.
* test-suite/tests/goops.test ("#:class slot allocation"): Add test.
| Andy Wingo | 2017-03-01 | 1 | -1/+1 |
| * | Flush when getting string from r6rs string output port...* module/rnrs/io/ports.scm (open-string-output-port): Calling the
get-string proc should flush the buffer and reset the file position.
* test-suite/tests/r6rs-ports.test ("8.2.10 Output ports"): Add tests.
Thanks to Freja Nordsiek for the report.
| Andy Wingo | 2017-03-01 | 1 | -1/+5 |
| * | Favor docstrings for describing the purpose of functions....* module/ice-9/boot-9.scm: Where functions have docstring-style
comments, make them proper docstrings.
| Wilfred Hughes | 2017-03-01 | 1 | -164/+104 |
| * | Better errors for odd-length keyword args...* libguile/vm-engine.c (bind-kwargs):
* libguile/vm.c (vm_error_kwargs_missing_value):
* libguile/eval.c (error_missing_value)
(prepare_boot_closure_env_for_apply): Adapt to mirror VM behavior.
* libguile/keywords.c (scm_c_bind_keyword_arguments): Likewise.
* module/ice-9/eval.scm (primitive-eval): Update to error on (foo #:kw)
with a "Keyword argument has no value" instead of the horrible "odd
argument list length". Also adapts to the expected args format for
the keyword-argument-error exception printer in all cases. Matches
1.8 optargs behavior also.
* test-suite/standalone/test-scm-c-bind-keyword-arguments.c (test_missing_value):
(missing_value_error_handler): Update test.
* test-suite/tests/optargs.test: Add tests.
| Andy Wingo | 2017-02-28 | 1 | -20/+31 |
| * | Declare module exports before loading imports...* module/ice-9/boot-9.scm (define-module*): Process module imports after
module exports. Allows for an additional kind of circular module
imports (see https://bugs.gnu.org/15540).
* test-suite/tests/modules.test ("circular imports"): Add test.
| Andy Wingo | 2017-02-28 | 1 | -40/+33 |
| * | Revert "futures: Limit the number of nested futures on the same stack."...This reverts commit 8a177d316c0062afe74f9a761ef460e297435e59, though
keeping the additional tests. (Guile 2.2 doesn't have a fixed stack
limit).
| Andy Wingo | 2017-02-28 | 1 | -16/+7 |
| * | Fix guild compile --to=cps / --from=cps...* module/language/cps/spec.scm (read-cps, write-cps): Fix CPS
serialization and parsing, so that "guild compile" works with --to=cps
and --from=cps.
| Andy Wingo | 2017-02-23 | 1 | -3/+17 |
| * | Fix kfun parse-cps bug...* module/language/cps.scm (parse-cps): Fix bug parsing kfun.
| Andy Wingo | 2017-02-23 | 1 | -1/+1 |
| * | Fix multi-arity dispatch in GOOPS...* module/oop/goops.scm (multiple-arity-dispatcher): Fix dispatch for
max-arity+1 when a generic is already in multiple-arity dispatch.
Fixes #24454.
* test-suite/tests/goops.test ("dispatch"): Add test.
| Andy Wingo | 2017-02-22 | 1 | -1/+1 |
| * | Fix compute-frame-sizes for case-lambda jumps...* module/system/vm/frame.scm (compute-frame-sizes): Fix for jumps to the
next arity.
| Andy Wingo | 2017-02-22 | 1 | -0/+3 |
| * | Fix flonum/complex type inference....* module/language/cps/types.scm (define-binary-result!): Arithmetic
where one argument is a flonum may produce a complex.
* test-suite/tests/compiler.test: Add test.
| Andy Wingo | 2017-02-19 | 1 | -2/+6 |
| * | Stringbufs immutable by default...* libguile/snarf.h (SCM_IMMUTABLE_STRINGBUF): Remove shared flag.
Stringbufs are immutable by default.
* libguile/strings.c: Rewrite blurb. Change to have stringbufs be
immutable by default and mutable only when marked as such. Going
mutable means making a private copy.
(STRINGBUF_MUTABLE, STRINGBUF_F_MUTABLE): New definitions.
(SET_STRINGBUF_SHARED): Remove.
(scm_i_print_stringbuf): Simplify to just alias the stringbuf as-is.
(substring_with_immutable_stringbuf): New helper.
(scm_i_substring, scm_i_substring_read_only, scm_i_substring_copy):
use new helper.
(scm_i_string_ensure_mutable_x): New helper.
(scm_i_substring_shared): Use scm_i_string_ensure_mutable_x.
(stringbuf_write_mutex): Remove; yaaaaaaaay.
(scm_i_string_start_writing): Use scm_i_string_ensure_mutable_x. No
more mutex.
(scm_i_string_stop_writing): Now a no-op.
(scm_i_make_symbol): Use substring/copy.
(scm_sys_string_dump, scm_sys_symbol_dump): Update.
* libguile/strings.h (SCM_I_STRINGBUF_F_SHARED): Remove.
(SCM_I_STRINGBUF_F_MUTABLE): Add.
* module/system/vm/assembler.scm (link-data): Don't add shared flag any
more. Existing compiled flags are harmless tho.
* test-suite/tests/strings.test ("string internals"): Update.
| Andy Wingo | 2017-02-16 | 1 | -5/+2 |
| * | Add accept4 support...* doc/ref/posix.texi (Network Sockets and Communication): Add
documentation.
* libguile/socket.c (scm_accept4): New function, replaces accept
implementation.
(scm_accept): Call scm_accept4.
(scm_init_socket): Define SOCK_CLOEXEC and SOCK_NONBLOCK.
* libguile/socket.h: Add private scm_accept4 decl.
* module/ice-9/suspendable-ports.scm (accept): Update.
| Andy Wingo | 2017-02-15 | 1 | -2/+2 |
| * | Web server and client use suspendable ports...* module/web/request.scm (write-request):
* module/web/response.scm (write-response): Use put-string instead of
display.
| Andy Wingo | 2017-02-08 | 2 | -2/+4 |
| * | Prevent TOCTTOU bugs in suspendable ports...* module/ice-9/suspendable-ports.scm: Prevent TOCTTOU bugs by
additionally returning the buffer and offset when we compute an
amount-buffered.
| Andy Wingo | 2017-02-08 | 1 | -81/+86 |
| * | Remove remaining "display" uses in (web http)...* module/web/http.scm (header-writer): Default to calling put-string.
(put-list): Rename from write-list, take the port first, and call the
put-item function with port then value. Adapt all callers.
(write-date): Rename display-digits to put-digits.
(put-challenge): Rename from write-challenge, adapt arguments to put
convention, and adapt callers.
(declare-symbol-list-header!): Use put-symbol.
(declare-integer-header!): Use put-non-negative-integer.o
(declare-entity-tag-list-header!): Use put-entity-tag-list.
("If-Range", "Etag"): Adapt to put-entity-tag.
(make-chunked-output-port): Use put-char.
| Andy Wingo | 2017-02-08 | 1 | -59/+62 |
| * | Beginnings of suspendable HTTP...* module/web/http.scm: Use put-string and other routines from (ice-9
textual-ports) in preference to `display'. The goal is for these
operations to be suspendable.
| Andy Wingo | 2017-02-08 | 1 | -134/+146 |
| * | Modernize (web http) a bit...* module/web/http.scm: Modernize the Guile Scheme by using more match,
when, unless, and non-tail conversion. No functional change, with the
exception of fixing a bug in write-key-value-list for symbols like
100-continue that shouldn't print as #{100-continue}#.
* test-suite/tests/web-http.test (pass-if-only-parse):
(pass-if-reparse, pass-if-parse): Arrange to also serialize and
reparse values from pass-if-parse. Apply to all existing tests except
fragments where we don't expect fragments to be written out.
| Andy Wingo | 2017-02-08 | 1 | -353/+356 |
| * | Fix rank 0 arrays and nested arrays in truncated-print...* module/ice-9/pretty-print.scm (print): In the array case, pass
#:inner? along to (print-sequence), unless we're at the last dimension
of the array.
Special case for 0-rank arrays, which cannot be empty and have no
length.
* test-suite/tests/print.test: Test some of the cases fixed by this
patch.
| Daniel Llorens | 2017-02-07 | 1 | -14/+21 |
