summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
* load: scope change for load hooks.Kaz Kylheku2021-09-032-20/+16
| | | | | | | | | | | | | | * eval.c (run_load_hooks): Install the given environment as dyn_env temporarily, and don't restore it until after calling the hooks. Thus the specified environment is now in effect when running the hooks. Also, pass nil to lookup_var, in the spirit of the previous commit. The fact that load_dyn_env had to be passed to lookup_var previously, which is an anti-pattern, tells us that this scoping rule was a code smell. If the *load-hooks* value comes from a given dynamic environment, then those functions should be executed in exactly that environment. * txr.1: Documentation updated.
* lookup_var: don't pass dyn_env explicitly.Kaz Kylheku2021-09-033-6/+6
| | | | | | | | | | | | | Since lookup_var(nil, ...) skips the environment and goes for dyn_env, there is no need to pass dyn_env explicitly; it's a bit of an anti-pattern. The argument is intended for lexical scopes. * eval.c (eval_exception, expand_eval, load): Pass nil to lookup_var instead of the current dynamic environment. * match.c (v_load): Likewise. * parser.c (txr_parse, read_eval_ret_last): Likewise.
* load: new macros push-after-load and pop-after-load.Kaz Kylheku2021-09-034-0/+73
| | | | | | | | | | | | | * eval.c (me_push_after_load, me_pop_after_load): New static functions. (eval_init): Register push-after-load and pop-after-load intrinsic macros. * tests/019/load-hook.tl: Tests for correct expansion. * txr.1: Documented. * stdlib/doc-syms.tl: Updated.
* New function: delcons.Kaz Kylheku2021-09-026-0/+110
| | | | | | | | | | | | * eval.c (eval_init): Register delcons intrinsic. * lib.[ch] (delcons): New function. * tests/010/cons.tl: New file. * txr.1: Documented. * stdlib/doc-syms.tl: Updated.
* load: new *load-hooks* feature.Kaz Kylheku2021-09-028-5/+136
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | *load-hooks* lets a .txr, .tl or .tlo file specify actions to be taken when the loading of that file completes, whether normally or via an exception. They are also honored by process exit. For instance, with this, we can have a Lisp file that behaves like a script which cleans up after itself (e.g. removing temporary files) even if it is not run as a stand-alone program, but invoked via (load ...). Because it's not a stand-alone program, it cannot simply use the at-exit-call mechanism. The unwind-protect operator could be used, but it's inconvenient because it protects a single form. The *load-hooks* feature in effect protects all the top level forms of a load, similarly to unwind-protect. Also, unwind-protect does not guard against a process exit. (However, *load-hooks* does not guard against an abnormal exit, only normal termination). * eval.c (load_hooks_s): New symbol variable. (run_load_hooks): New function. (run_load_hooks_atexit): New static function. (load): bind *load-hooks* to nil around load. Implement the hooks processing via run_load_hooks, taking care to pass the load-time dynamic environment that has already been undone. (eval_init): Initialize load_hooks_s and register the *load-hooks* variable. Register run_load_hooks_atexit with atexit, so the current value of *load-hooks* is processed on process exit. * eval.h (load_hooks_s, run_load_hooks): Declared. * match.c (v_load): Similar changes as in load. * txr.c (txr_main): Run the load hooks with run_load_hooks immediately after processing the .txr or .tl file, before entering the listener. * tests/019/load-hook.tl: New directory and file * tests/load-hook.tl: New file. * txr.1: Documented. * stdlib/doc-syms.tl: Updated.
* tags: process only files in repo.Kaz Kylheku2021-09-011-1/+2
| | | | | | | | | * libtags.txr: use git ls-files instead of glob to obtain list of .c files. This also means that we go into subdirectories now, since git ls-files '*.c' lists items like linenoise/linenoise.c and chksums/md5.c. For now, we are not finding any new tags in these places, but in the future that could change.
* compiler: fix output path of compiled files.Paul A. Patience2021-09-011-1/+1
| | | | | | | | | | | The open-compile-streams function was calling trim-right with the arguments in the wrong order, resulting in an output path equal to the suffix of the input path. Regression introduced in 8d8fee2e506806d9c117b17432ef3a5ec0d6f457. * stdlib/compiler.tl (open-compile-streams): Swap in-path and suff arguments in trim-right call.
* tags: fix warnings using @(mdo ...).Kaz Kylheku2021-08-311-10/+10
| | | | | | | * libtags.txr: Use @(mdo) for defining variables and structures, so this is done as libtags.txr is parsed. The remaining top-level actions are split off in a separate @(do ...) placed later.
* tags: add script to tag interned symbols.Paul A. Patience2021-08-312-82/+610
| | | | * libtags.txr: New file. This is a work in progress.
* Version 270.txr-270Kaz Kylheku2021-08-304-5/+28
| | | | | | | | | | * RELNOTES: Updated. * configure (txr_ver): Bumped version. * stdlib/ver.tl (lib-version): Bumped. * txr.1: Bumped version and date.
* seq_iter: some new test cases.Kaz Kylheku2021-08-301-0/+8
| | | | | * tests/012/iter.tl: Several new cases to provide some coverage in recently fixed areas. All of these break in 268.
* doc: fix reference to buf-sub.Paul A. Patience2021-08-301-2/+2
| | | | | | It should be sub-buf. * txr.1: buf-sub -> sub-buf.
* doc: fix incorrect slet description.Paul A. Patience2021-08-291-1/+1
| | | | | | slet is a stronger form of rlet, not weaker. * txr.1: weaker -> stronger.
* buf, carray: remove de trop binding in accessors.Paul A. Patience2021-08-291-6/+2
| | | | | | | | They were forgotten in commit 1fb6f6691d5b6fb6b037bb14073694f651f2b9fc. * stdlib/ffi.tl (carray-sub, sub-buf): Remove binding which ensures the new value is evaluated only once (defset does this already).
* Makefile: whitespace fix.Kaz Kylheku2021-08-291-3/+3
| | | | | * Makefile (SRCS): Fix mixture of tabs and spaces, and bad alignment, reported by Paul A. Patience.
* open-file: improvement: "a" mode sets create flag.Kaz Kylheku2021-08-291-2/+2
| | | | | | | * stream.c (w_fopen_mode): Only test the m.create flag as the basis for O_CREAT, not m.append. (do_parse_mode): In the 'a' case, set m.create = 1, since all variants of append mode create the file.
* open-file: add mode option "x".Paul A. Patience2021-08-293-5/+26
| | | | | | | | | | | | * stream.h (struct stdio_mode): New member, excl flag. (stdio_mode_init_blank, stdio_mode_init_r, stdio_mode_init_rpb): Add initializer for excl flag. * stream.c (do_parse_mode): Handle 'x' in mode string. (w_fopen_mode): Add O_EXCL flag if m.excl is set. Throw an error if we don't HAVE_FCNTL and m.excl is set. * txr.1: Document mode option "x".
* open-file: fix broken file-creation modes.Paul A. Patience2021-08-291-2/+3
| | | | | | | | The "w+", "m+" and "a+" modes wouldn't create the file. * stream.c (w_fopen_mode): Add O_TRUNC and O_CREAT flags if m.create or m.append is set, rather than if m.read is unset and m.write is set.
* seq_iter: allow mixed fixnum/bignum ranges.Kaz Kylheku2021-08-291-6/+18
| | | | | | * lib.c (seq_iter_init_with_info): The to value in a range could be a bignum, which we should treat as a bignum range, rather than blowing up due to calling c_num on that value.
* seq_iter: fix gc issues.Kaz Kylheku2021-08-292-26/+25
| | | | | | | | | | | | | | | | | | | | * lib.h (struct seq_iter_ops): New operation, mark. (seq_iter_ops_init): Default the mark operation to seq_iter_mark_op. (seq_iter_ops_init_nomark): New macro. * lib.c (seq_iter_mark_op): New static function. (si_range_cnum_ops, si_range_chr_ops, si_rev_range_cnum_ops, si_rev_range_chr_ops, si_chr_ops): Initialize with seq_iter_ops_init_nomark so that the iterator has no mark operation. All other iterator types have the above new static function as their mark op. (seq_iter_mark): Simplified: if the iterator has a mark op, call it. (seq_next, iter_step): If the iterator has a mark op, that means that the seq_get operation is likely replacing one heap object with another, and so mut(iter) must be called to inform the garbage collector of the assignment.
* seq_iter: refactoring.Kaz Kylheku2021-08-292-42/+88
| | | | | | | | | | | | | | | | | | | | | This is needed in preparation for fixing some gc bugs in iteration in a less hacky way. * lib.h (struct seq_iter): Members get and peek removed, replaced by ops pointer. (struct seq_iter_ops): New struct type. (seq_get, seq_peek): Call get and peek through ops table. * lib.c (seq_geti): Refer to get through ops. (si_null_ops, si_list_ops, si_vec_ops, si_hash_ops, si_tree_ops, si_range_cnum_ops, si_range_chr_ops, si_range_bignum_ops, si_range_str_ops, si_rev_range_cnum_ops, si_rev_range_chr_ops, si_rev_range_bignum_ops, si_rev_range_str_ops, si_chr_ops, si_num_ops, si_oop_ops, si_fast_oop_ops): New static structures. (seq_iter_init_with_info): Point new ops member of seq_iter to one of the new structures. Remove initializations of get and peek.
* Version 269.txr-269Kaz Kylheku2021-08-287-1121/+1193
| | | | | | | | | | | | | | * RELNOTES: Updated. * configure (txr_ver): Bumped version. * stdlib/ver.tl (lib-version): Bumped. * txr.1: Bumped version and date. * txr.vim, tl.vim: Regenerated. * protsym.c: Likewise.
* doc: problem with Unix Memory Mapping heading.Kaz Kylheku2021-08-281-1/+1
| | | | * txr.1: fix .SS to .SS*.
* open-file: fix broken "+" mode string.Paul A. Patience2021-08-281-2/+1
| | | | | | | | The "+" mode string should be equivalent to "r+", according to the manual, but before this change it was equivalent to "r". * stream.c (do_parse_mode): Unconditionally set m.write to 1 when "+" is present in the mode string.
* mmap test: use random-buf.Kaz Kylheku2021-08-261-4/+2
| | | | | * tests/017/mmap.tl: Use the new random-buf function instead of converting pseudo-random bignums to a buffer.
* rand: provide random-buf function.Kaz Kylheku2021-08-264-0/+67
| | | | | | | | | | | * rand.c (random_buf): New function. (rand-init): random-buf intrinsic registered. * rand.h (random_buf): Declared. * txr.1: Documented. * stdlib/doc-syms.tl: Updated.
* itypes: provide a c_size function for size_t.Kaz Kylheku2021-08-262-0/+17
| | | | * itypes.[ch] (c_size): New function.
* random: bug: 32 bit shift.Kaz Kylheku2021-08-261-3/+3
| | | | | | | | | | | * rand.c (random): The msb_rand_mask calculation shifts by 32 when msb_rand_bits is zero, like when the modulus is (expt 2 32). By fluke, this has been treated as a zero bit shift on 64 bit intel, so the correct mask was calculated. However, in a PPC64 build, it yields zero. Instead, we calculate the required shift by negating the bit count, in unsigned semantics, and then reducing modulo 32. This gives us the correct value from 0 to 31, resulting in the correct mask.
* mmap test: bug: random buffer generation.Kaz Kylheku2021-08-261-1/+2
| | | | | | | * tests/017/mmap.tl: The random buffer procedure can generate buffers which are too small, because the random integer values are anywhere from 0 to the modulus. This showed up up on PPC-64 with 65536 byte pages.
* ffi: improve support for big/little-endian types.Kaz Kylheku2021-08-261-232/+187
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Instead of byte reads/writes, we use byte order swapping. Reads and writes that are aligned take place as a single data transfer, which makes it possible to use these be/le types for accessing hardware registers. This is useful in systems where the host processor accesses the bus in opposite endian relative to some peripheral. Moreover, when the functions match the local endian, we just use the native get/put functions via #define macro redirection. The source code size is reduced. * ffi.c (ffi_swap_u16, ffi_swap_u32, ffi_swap_u64, ffi_swap_i16, ffi_swap_i32, ffi_swap_i64): New static functions. (ffi_swap_i16_put, ffi_swap_i16_get, ffi_swap_u16_put, ffi_swap_u16_get, ffi_swap_i32_put, ffi_swap_i32_get, ffi_swap_u32_put, ffi_swap_u32_get, ffi_swap_i64_put, ffi_swap_i64_get, ffi_swap_u64_put, ffi_swap_u64_get): New static functions. (ffi_be_i16_put, ffi_be_i16_get, ffi_be_u16_put, ffi_be_u16_get, ffi_be_i32_put, ffi_be_i32_get, ffi_be_u32_put, ffi_be_u32_get, ffi_be_i64_put, ffi_be_i64_get, ffi_be_u64_put, ffi_be_u64_get, ffi_le_i16_put, ffi_le_i16_get, ffi_le_u16_put, ffi_le_u16_get, ffi_le_i32_put, ffi_le_i32_get, ffi_le_u32_put, ffi_le_u32_get, ffi_le_i64_put, ffi_le_i64_get, ffi_le_u64_put, ffi_le_u64_get): Functions deleted, replaced by same-named #define macros to redirect to the native functions or the _swap_ functions based on the endian. (ffi_be_i16_rput, ffi_be_i16_rget, ffi_be_u16_rput, ffi_be_u16_rget, ffi_be_i32_rput, ffi_be_i32_rget, ffi_be_u32_rput, ffi_be_u32_rget, ffi_le_i16_rput, ffi_le_i16_rget, ffi_le_u16_rput, ffi_le_u16_rget, ffi_le_i32_rput, ffi_le_i32_rget, ffi_le_u32_rput, ffi_le_u32_rget): Functions wrapped with #if in case these exact width types don't exist. (ffi_init_types): Wrap some exact-width-type type definitions with #if in case the types don't exist.
* solaris: disable failing test case.Kaz Kylheku2021-08-261-0/+4
| | | | | | | * tests/007/except-4.txr: This new test case does not work on Solaris 10 because a shell script that kills itself via kill $$ appears to be terminating successfully with an exit status of 208, not appearing to be killed by a signal.
* mmap: Solaris 10 build fix.Kaz Kylheku2021-08-251-0/+20
| | | | | | | | | | | | | | Solaris 10's <sys/mman.h> header doesn't declare madvise if _POSIX_C_SOURCE > 2 is defined or _XPTG4_2, even though __EXTENSIONS__ is defined. This is a bug in the header. * configure (solaris_target): New variable. We set this to y in several places where the script discovers that -D__EXTENSIONS__ is required to get something to work, which is a Solaris 10 telltale sign. Then just before the mmap test, if solaris_target is true, we deposit a blurb into config.h that declares madvise. This way, the ffi.c code, or any other committed C code doesn't have to carry any Solaris hack.
* mmap tests: BSD patch.Kaz Kylheku2021-08-251-1/+2
| | | | | | * tests/017/mmap.tl: On BSD, the map-anon test case where we don't specify map-private or map-shared doesn't result in an invalid argument error; a mapping is produced.
* doc: document daemon function arguments.Kaz Kylheku2021-08-251-1/+7
| | | | | * txr.1: Fix the neglect to specify that the daemon function's arguments are Boolean values, rather than integers.
* doc: improvement in awk range operators.Kaz Kylheku2021-08-251-8/+20
| | | | | | * txr.1: Prompted by a sentence with faulty grammar, improving the description of the table which summarizes the semantics of the nine range operators.
* carray: allow displacement in carray-pun.Kaz Kylheku2021-08-253-12/+63
| | | | | | | | | | * ffi.c (carray_pun): Takes two new arguments to specify displacement and size. (ffi_init): Registration of carray-pun adjusted. * ffi.h (carray_pun): Declaration updated. * txr.1: Documented.
* ffi: provide mmap through carray.Kaz Kylheku2021-08-226-0/+953
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * configure: configure test for mmap depositing HAVE_MMAP into config.h. * ffi.c (struct carray): Subject to HAVE_MMAP, new mm_len member which keeps track of the size of an underlying mapping so that we can unmap it, as well as peform operations like msync on it. (make_carray): Initialize mm_len to 0. (MAP_GROWSDOWN, MAP_LOCKED, MAP_NORESERVE, MAP_POPULATE, MAP_NONBLOCK, MAP_STACK, MAP_HUGETLB, MAP_SHARED, MAP_PRIVATE, MAP_FIXED, MAP_ANON, MAP_HUGE_SHIFT, MAP_HUGE_MASK, PROT_READ, PROT_WRITE, PROT_EXEC, PROT_NONE, PROT_GROWSDOWN, PROT_GROWSUP, MADV_NORMAL, MADV_RANDOM, MADV_SEQUENTIAL, MADV_WILLNEED, MADV_DONTNEED, MADV_FREE, MADV_REMOVE, MADV_DONTFORK, MADV_DOFORK, MADV_MERGEABLE, MADV_UNMERGEABLE, MADV_HUGEPAGE, MADV_NOHUGEPAGE, MADV_DONTDUMP, MADV_DODUMP, MADV_WIPEONFORK, MADV_KEEPONFORK, MADV_HWPOISON, MS_ASYNC, MS_SYNC, MS_INVALIDATE): #define as 0 if missing. (carray_munmap_op): New static function. (carray_mmap_ops): New static structure. (mmap_wrap, munmap_wrap): New functions. (mmap_op): New static function. (mprotect_wrap, madvise_wrap, msync_wrap): New functions. (ffi_init): Register mmap, munmap, mprotect, madvise and msync as well as numerous integer variables: map-growsdown, map-locked, map-noreserve, map-populate, map-nonblock, map-stack, map-hugetlb, map-shared, map-private, map-fixed, map-anon, map-huge-shift, map-huge-mask, prot-read, prot-write, prot-exec, prot-none, prot-growsdown, prot-growsup, madv-normal, madv-random, madv-sequential, madv-willneed, madv-dontneed, madv-free, madv-remove, madv-dontfork, madv-dofork, madv-mergeable, madv-unmergeable, madv-hugepage, madv-nohugepage, madv-dontdump, madv-dodump, madv-wipeonfork, madv-keeponfork, madv-hwpoison, ms-async, ms-sync, ms-invalidate, page-size. * ffi.h (mmap_wrap, munmap_wrap, mprotect_wrap madvise_wrap, msync_wrap): Declared. * tests/017/mmap.tl: New file. * txr.1: Documented. * stdlib/doc-syms.tl: Updated.
* doc: misspelled cond in ecase section.Kaz Kylheku2021-08-221-1/+1
| | | | * txr.1: Fix cond to codn.
* sub: allow generic iterables.Kaz Kylheku2021-08-223-4/+109
| | | | | | | | | | | | | | | | In this patch we allow the s in (sub s from to) and [s from..to] to be any iterable. * lib.c (iter_dynamic, sub_iter): New static function. (generic_funcall): Handle all objects via the sequence case: ref, sub and all that. Unfortunately, we lose some error handling at the level of the sub function. But we allow any iterable to be passed through to sub. (sub): Handle default case through sub_iter. * tests/012/iter.tl: New cases. * txr.1: Documented.
* iter-begin: allow iterator argument.Kaz Kylheku2021-08-222-0/+16
| | | | | | | | * lib.c (seq_iter_init_with_info): Allow the iterated object to be an iterator, in which case a copy of the iterator is set up. * txr.1: Documented.
* sub-list: better handling of from value of t.Kaz Kylheku2021-08-221-4/+2
| | | | | | | | * lib.c (sub_list): If from is t, then just return nil. Do not reset it to nil in this case. After this we know from is not nil; we don't have to check for this inside one loop. That loop was wastefully iterating over the list in the from == nil case only to calculate nil.
* iter-begin: string range support.Kaz Kylheku2021-08-223-3/+178
| | | | | | | | | | | | | | Ranges like "AAA".."ZZZ" are now possible. * lib.c (seq_iter_get_range_str, seq_iter_peek_range_str, seq_iter_get_rev_range_str): New static functions. (seq_iter_init_with_info): Support string ranges via above new functions. Range direction test is now done with less and equal rather than lt and gt. * tests/012/iter.tl: New file. * txr.1: Documented.
* ecase: diagnose bad syntax.Kaz Kylheku2021-08-211-0/+4
| | | | * eval.c (me_ecase): Diagnose missing test form, like me_case.
* configure: implement full-repl option.Kaz Kylheku2021-08-205-39/+53
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch unbundles the building of the full-featured REPL from HAVE_TERMIOS. We make it subject to its own configuration option CONFIG_FULL_REPL, which is 1 by default. This way, the downstream users or package maintainers can build TXR without the full-featured REPL even if HAVE_TERMIOS is 1, and the other termios material is built-in. * configure (full_repl): New variable. (help): Include full-repl in the help text. In the termios test, if we don't detect termios, then negate the full_repl variable. In the final config variable generation section, generate the CONFIG_FULL_REPL 1 define in config.h, if full_repl is true, ensuring it is subject to HAVE_TERMIOS, too. * linenoise/linenoise.c: Replace HAVE_TERMIOS with CONFIG_FULL_REPL. * linenoise/linenoise.h: Likewise. * parser.c: Likewise. * txr.c: Likewise and ... (if_termios): Macro renamed to if_full_repl. (if_full_repl): New macro. (opt_noninteractive): Use if_full_repl macro for initializing. (banner): Use if_ful_repl macro instead of if_termios.
* listener: additional reductions in non-termios build.Kaz Kylheku2021-08-203-6/+53
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * linenoise/linenoise.c (struct lino_state): these members are now absent in non-termios build: atom_callback, ca_ctx, rawmode, mlmode, clip, result, buf, plen, pos, sel, end, len, dlen, dpos, dsel, dend, cols, oldrow, maxrows, history_index, need_resize, need_refresh, selmode, selinclusive, noninteractive and undo_stack. (lino_set_multiline, lino_get_multiline, lino_set_selinclusive, lino_get_selinculsive, lino_set_noninteractive, lino_get_noninteractive, lino_set_atom_cb): Functions now only defined in termios build. (linenoise): Adjustments for missing members in non-termios mode. (lino_make): We no longer need to set noninteractive to 1 in non-termios build. The flag no longer exists. (lino_copy, lino_cleanup): Avoid referencing nonexistent members in non-termios build. (lino_set_result): Another termios-only function. * linenoise/linenoise.h (lino_set_result, lino_clear_screen, lino_set_multiline, lino_get_multiline, lino_set_selinclusive, lino_get_selinculsive, lino_set_noninteractive, lino_get_noninteractive, lino_atom_cb_t, lino_set_atom_cb): Declare only in termios build. * parser.c (repl): Add #if HAVE_TERMIOS to avoid using linenoise features not available in non-termios build, and to remove any variables that thus become unused.
* listener: unbundle from termios.Kaz Kylheku2021-08-206-47/+59
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit fixes the conceptual issue that when there is no termios support (HAVE_TERMIOS is absent/false), then there is no listener at all, even though the listener supports plain mode that doesn't require termios. * Makefile (linenoise/linenoise.o): Link in unconditionally, not subject to have_termios. * linenoise.c: Include termios-related header only if HAVE_TERMIOS. (struct lino_state): Define the completion_callback and orig_termios members only if HAVE_TERMIOS. (wcnsprintf, atexit_handler, enable_raw_mode, disable_raw_mode, get_cursor_position, get_columns, lino_clear_screen, refresh_line, handle_resize, generate_beep, delete_undo, free_undo_stack, record_undo, record_triv_undo, remove_noop_undo, restore_undo, undo_subst_hist_idx, undo_renumber_hist_idx, free_completions, sync_data_to_buf, compare_completions, complete_line, lino_set_completion_cb, lino_add_completion, next_hist_match, copy_display_params, history_search, ab_init, ab_append, ab_free, sync_data_to_buf, copy_display_params, refresh_singleline, col_offset_in_str, refresh_multiline, refresh_line, move_cursor_multiline, move_cursor, scan_match_rev, scan_rev, scan_match_fwd, scan_fwd, find_nearest_paren, usec_delay, paren_jump, flash, yank, yank_by_ptr, update_sel, clear_sel, yank_sel, delete_sel, edit_insert, edit_insert_str, edit_move_left, edit_move_right, edit_move_home, edit_move_sol, edit_move_end, edit_move_eol, edit_move_matching_paren, edit_history_next, edit_delete, edit_backspace, edit_delete_prev_all, edit_delete_to_eol, edit_delete_prev_word, edit_delete_line, tr, char, edit_in_editor, edit, sigwinch_handler): Functions defined only if HAVE_TERMIOS. (struct abuf, struct row_values): Struct types defined only if HAVE_TERMIOS. (screen_rows): Defined only if HAVE_TERMIOS. (linenoise): Support only noninteractive read loop unless HAVE_TERMIOS. (lino_make): If HAVE_TERMIOS is false, then set the noninteractive flag, so the linenoise function enters the plain-mode loop. (lino_cleanup, lino_hist_add): Add #ifdefs to avoid calling nonexistent functions when HAVE_TERMIOS is false. * linenoise/linenoise.h (struct lino_completions, lino_compl_cb_t): Define these types only if HAVE_TERMIOS. (lino_set_completion_cb, lino_add_completion): Declare only if HAVE_TERMIOS. * parser.c: Include linenoise/linenoise.h unconditionally. (report_security_problem, load_rcfile, repl_intr, read_eval_ret_last, get_home_path, repl_warning, is_balanced_line, hist_save): Now define regardless of HAVE_TERMIOS. (repl): Define regardless of HAVE_TERMIOS, but don't set completion or atom callback if HAVE_TERMIOS is false. * parser.h (repl): Declare unconditionally, not subject to HAVE_TERMIOS. * txr.c (if_termios): New macro. (opt_noninteractive): Initialize to 1 if HAVE_TERMIOS is false. (help): Text about entering into listener mode is always present now, even in a build withou HAVE_TERMIOS. (banner): Function is always defined. If we don't HAVE_TERMIOS, then the unused string literal that will never be printed is replaced by nil. (hint): Function removed. (txr_main): Blocks conditional on HAVE_TERMIOS that either call banner and go to the repl, or else call hint and exit, are reduced to unconditionally calling banner and going to the repl. All #if HAVE_TERMIOS blocks are similarly replaced with just the HAVE_TERMIOS case.
* matcher: rename error-throwing macros.Kaz Kylheku2021-08-195-24/+24
| | | | | | | | | | | | | | | * stdlib/match.tl (must-match): Renamed to just match. It's just when-match without the "when". (must-match-case): Renamed to match-ecase, consistent with the case -> ecase naming scheme. * lisplib.c (match_set_entries): Names updated here. * tests/011/patmatch.tl: Test cases updated. * txr.1: Names updated here. * stdlib/doc-syms.tl: Updated.
* hash: use unsigned, and operation.Kaz Kylheku2021-08-192-16/+16
| | | | | | | | | | | | | | | | * hash.c (struct hash): modulus and count change from cnum to ucnum. (hash_mark, hash_grow, copy_hash, do_weak_tables): Use ucnum local vars. (do_make_hash, make_similar_hash): Use c_unum to obtain modulus. (gethash_c, gethash_e): Use & masking operation to reduce hash value to table size. (remhash): Move sanity check before decrement since unsigned value can't go below zero. (clearhash): Use ucnum and c_unum. (hash_iter_peek): Use ucnum for chain count local. * hash.h (struct hash_iter): chain changes from cnum to ucnum.
* New ecase macros.Kaz Kylheku2021-08-183-1/+116
| | | | | | | | | | | | | | | | | | | | Even prior to discovering the recent defect in deffi, which was caused by a missing case in caseql, combined with poor testing, I was already thinking about adding ecase macros. The introduction of must-match and must-match-case also shows my motivation. That deffi bug convinced me to take action and implement these. * eval.c (case_error_s) New symbol variable. (me_ecase): New static function. (eval_init): Register new intrinsic macros ecaseq, ecaseql, ecasequal, ecaseq*, ecaseql* and ecasequal*. Intern case-error and initialize case_error_s. * txr.1: Documented. Also updated Exception Hierarchy diagram with match-error and case-error. * stdlib/doc-syms.tl: Updated.
* eval: de-duplicate built-in macro definitions.Kaz Kylheku2021-08-181-19/+26
| | | | | | | * eval.c (eval_init): Numerous macros share the same implementation function, and their registrations make wasteful repeated func_f2 calls to hoist that function from C to Lisp more than once. Let's go through and condense all of them.
generated by cgit v1.2.3 (git 2.25.1) at 2025-03-17 16:14:16 +0000