Class: VibeZstd::DCtx

Inherits:
Object
  • Object
show all
Defined in:
ext/vibe_zstd/vibe_zstd.c

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.default_initial_capacity ⇒ Object

DCtx default_initial_capacity getter (class method)



181
182
183
184
185
186
187
 
# File 'ext/vibe_zstd/dctx.c', line 181

static VALUE
vibe_zstd_dctx_get_default_initial_capacity(VALUE self) {
    if (default_initial_capacity == 0) {
        return SIZET2NUM(ZSTD_DStreamOutSize());
    }
    return SIZET2NUM(default_initial_capacity);
}

.default_initial_capacity=(value) ⇒ Object

DCtx default_initial_capacity setter (class method)



190
191
192
193
194
195
196
197
198
199
200
201
202
 
# File 'ext/vibe_zstd/dctx.c', line 190

static VALUE
vibe_zstd_dctx_set_default_initial_capacity(VALUE self, VALUE value) {
    if (NIL_P(value)) {
        default_initial_capacity = 0;  // Reset to default
    } else {
        size_t capacity = NUM2SIZET(value);
        if (capacity == 0) {
            rb_raise(rb_eArgError, "initial_capacity must be positive (or nil to reset to default)");
        }
        default_initial_capacity = capacity;
    }
    return value;
}

.default_max_decompressed_size ⇒ Object

DCtx default_max_decompressed_size getter (class method); 0 = unlimited



236
237
238
239
 
# File 'ext/vibe_zstd/dctx.c', line 236

static VALUE
vibe_zstd_dctx_get_default_max_decompressed_size(VALUE self) {
    return SIZET2NUM(default_max_decompressed_size);
}

.default_max_decompressed_size=(value) ⇒ Object

DCtx default_max_decompressed_size setter (class method)



242
243
244
245
246
247
248
249
250
 
# File 'ext/vibe_zstd/dctx.c', line 242

static VALUE
vibe_zstd_dctx_set_default_max_decompressed_size(VALUE self, VALUE value) {
    if (NIL_P(value)) {
        default_max_decompressed_size = 0;  // unlimited
    } else {
        default_max_decompressed_size = NUM2SIZET(value);
    }
    return value;
}

.estimate_memory ⇒ Object

DCtx.estimate_memory()



55
56
57
58
59
 
# File 'ext/vibe_zstd/dctx.c', line 55

static VALUE
vibe_zstd_dctx_estimate_memory(VALUE self) {
    size_t estimate = ZSTD_estimateDCtxSize();
    return SIZET2NUM(estimate);
}

.frame_content_size(data) ⇒ Object

DCtx frame_content_size - class method to get frame content size



461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
 
# File 'ext/vibe_zstd/dctx.c', line 461

static VALUE
vibe_zstd_dctx_frame_content_size(VALUE self, VALUE data) {
    StringValue(data);
    unsigned long long contentSize = ZSTD_getFrameContentSize(RSTRING_PTR(data), RSTRING_LEN(data));

    if (contentSize == ZSTD_CONTENTSIZE_ERROR) {
        return Qnil;  // Invalid frame
    }

    if (contentSize == ZSTD_CONTENTSIZE_UNKNOWN) {
        return Qnil;  // Unknown size
    }

    return ULL2NUM(contentSize);
}

.parameter_bounds ⇒ Object

Instance Method Details

#decompress(*args) ⇒ Object

Skippable frames at the beginning of data are automatically skipped.



489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
 
# File 'ext/vibe_zstd/dctx.c', line 489

static VALUE
vibe_zstd_dctx_decompress(int argc, VALUE* argv, VALUE self) {
    VALUE data, options = Qnil;
    rb_scan_args(argc, argv, "1:", &data, &options);
    vibe_zstd_dctx* dctx;
    TypedData_Get_Struct(self, vibe_zstd_dctx, &vibe_zstd_dctx_type, dctx);
    StringValue(data);
    const char* src = RSTRING_PTR(data);
    size_t srcSize = RSTRING_LEN(data);
    size_t offset = 0;

    // Magicless frames (format = ZSTD_f_zstd1_magicless) carry no magic number,
    // so frame introspection (content size, dict ID, skippable detection) cannot
    // be performed. Force the streaming decompress path, which honors the format
    // parameter set on the context via ZSTD_decompressStream.
    int dformat = 0;
    (void)ZSTD_DCtx_getParameter(dctx->dctx, ZSTD_d_format, &dformat);
    int magicless = (dformat == ZSTD_f_zstd1_magicless);

    unsigned long long contentSize;
    unsigned int frame_dict_id;

    if (magicless) {
        contentSize = ZSTD_CONTENTSIZE_UNKNOWN;  // route to streaming path
        frame_dict_id = 0;                        // cannot read dict ID without magic
    } else {
        // Skip any leading skippable frames
        while (offset < srcSize && ZSTD_isSkippableFrame(src + offset, srcSize - offset)) {
            size_t frameSize = ZSTD_findFrameCompressedSize(src + offset, srcSize - offset);
            if (ZSTD_isError(frameSize)) {
                rb_raise(rb_eRuntimeError, "Invalid skippable frame at offset %zu: %s", offset, ZSTD_getErrorName(frameSize));
            }
            offset += frameSize;
        }

        // Now check the actual compressed frame
        if (offset >= srcSize) {
            rb_raise(rb_eRuntimeError, "No compressed frame found in %zu bytes (only skippable frames)", srcSize);
        }

        src += offset;
        srcSize -= offset;

        contentSize = ZSTD_getFrameContentSize(src, srcSize);
        if (contentSize == ZSTD_CONTENTSIZE_ERROR) {
            rb_raise(rb_eRuntimeError, "Invalid compressed data: not a valid zstd frame (size: %zu bytes)", srcSize);
        }

        // Check dictionary requirements from the frame
        frame_dict_id = ZSTD_getDictID_fromFrame(src, srcSize);
    }

    // Extract keyword arguments
    ZSTD_DDict* ddict = NULL;
    unsigned int provided_dict_id = 0;
    size_t initial_capacity = 0;  // 0 = not specified in per-call options
    size_t max_size = 0;          // 0 = not specified in per-call options

    if (!NIL_P(options)) {
        VALUE dict_val = rb_hash_aref(options, ID2SYM(rb_intern("dict")));
        if (!NIL_P(dict_val)) {
            vibe_zstd_ddict* ddict_struct;
            TypedData_Get_Struct(dict_val, vibe_zstd_ddict, &vibe_zstd_ddict_type, ddict_struct);
            ddict = ddict_struct->ddict;
            provided_dict_id = ZSTD_getDictID_fromDDict(ddict);
        }

        VALUE initial_capacity_val = rb_hash_aref(options, ID2SYM(rb_intern("initial_capacity")));
        if (!NIL_P(initial_capacity_val)) {
            initial_capacity = NUM2SIZET(initial_capacity_val);
            if (initial_capacity == 0) {
                rb_raise(rb_eArgError, "initial_capacity must be positive");
            }
        }

        // Per-call output-size limit; accepts :max_decompressed_size or :max_size.
        VALUE max_size_val = rb_hash_aref(options, ID2SYM(rb_intern("max_decompressed_size")));
        if (NIL_P(max_size_val)) {
            max_size_val = rb_hash_aref(options, ID2SYM(rb_intern("max_size")));
        }
        if (!NIL_P(max_size_val)) {
            max_size = NUM2SIZET(max_size_val);
            if (max_size == 0) {
                rb_raise(rb_eArgError, "max_decompressed_size must be positive");
            }
        }
    }

    // Resolve max_size fallback chain: per-call > instance > class default.
    // A value of 0 at every level means unlimited.
    if (max_size == 0) {
        max_size = dctx->max_decompressed_size;  // instance
        if (max_size == 0) {
            max_size = default_max_decompressed_size;  // class
        }
    }

    // Validate dictionary matches frame requirements
    if (frame_dict_id != 0 && ddict == NULL) {
        rb_raise(rb_eArgError, "Data requires dictionary (dict_id: %u) but none provided", frame_dict_id);
    }

    if (ddict != NULL && frame_dict_id != 0 && provided_dict_id != frame_dict_id) {
        rb_raise(rb_eArgError, "Dictionary mismatch: frame requires dict_id %u, provided dict_id %u",
                 frame_dict_id, provided_dict_id);
    }

    // Resolve initial_capacity fallback chain: per-call > instance > class default > ZSTD default
    if (initial_capacity == 0) {
        initial_capacity = dctx->initial_capacity;  // Instance default
        if (initial_capacity == 0) {
            initial_capacity = default_initial_capacity;  // Class default
            if (initial_capacity == 0) {
                initial_capacity = ZSTD_DStreamOutSize();  // ZSTD default (~128KB)
            }
        }
    }

    // If content size is unknown, use streaming decompression with exponential growth.
    // Releases GVL to allow other Ruby threads to run during decompression.
    // Uses C malloc/realloc (not Ruby allocators) since Ruby API calls are forbidden without GVL.
    if (contentSize == ZSTD_CONTENTSIZE_UNKNOWN) {
        // Reference the dictionary on the context before streaming decompression.
        // ZSTD_decompressStream uses whatever dict is referenced on the DCtx, so
        // without this the dictionary would be ignored on the unknown-size path
        // (every dict frame produced by CompressWriter has unknown content size).
        if (ddict) {
            size_t rd = ZSTD_DCtx_refDDict(dctx->dctx, ddict);
            if (ZSTD_isError(rd)) {
                rb_raise(rb_eRuntimeError, "Failed to reference dictionary: %s", ZSTD_getErrorName(rd));
            }
        }

        decompress_stream_nogvl_args stream_args = {
            .dctx = dctx->dctx,
            .src = src,
            .src_size = srcSize,
            .dst = NULL,
            .dst_capacity = 0,
            .dst_size = 0,
            .initial_capacity = initial_capacity,
            .max_size = max_size,
            .error = 0,
            .limit_exceeded = 0,
            .truncated = 0,
            .error_name = NULL
        };

        // Run the streaming decompression and build the result under rb_ensure:
        // the cleanup frees the C buffer and un-references the dictionary on
        // every exit path, including the raises below and async exceptions
        // delivered when the GVL is reacquired.
        dctx_stream_decompress_state state = {
            .dctx = dctx->dctx,
            .ddict = ddict,
            .args = &stream_args,
            .data = data,
            .max_size = max_size
        };
        return rb_ensure(vibe_zstd_dctx_stream_decompress_body, (VALUE)&state,
                         vibe_zstd_dctx_stream_decompress_cleanup, (VALUE)&state);
    }
    // Reject a frame whose declared content size exceeds the limit before
    // allocating the output buffer (the header is attacker-controlled).
    if (max_size && contentSize > (unsigned long long)max_size) {
        rb_raise(rb_eDecompressedSizeExceeded,
                 "Declared content size %llu exceeds limit of %zu bytes", contentSize, max_size);
    }

    VALUE result = rb_str_new(NULL, contentSize);
    decompress_args args = {
        .dctx = dctx->dctx,
        .ddict = ddict,
        .src = src,
        .srcSize = srcSize,
        .dst = RSTRING_PTR(result),
        .dstCapacity = contentSize,
        .result = 0
    };
    // Lock the source string while the GVL is released: another Ruby thread
    // holding the same string must not mutate or GC it mid-decompression.
    // The helper unlocks via rb_ensure so an async exception cannot leave
    // the string permanently locked.
    vibe_zstd_nogvl_with_str_locked(decompress_without_gvl, &args, data);
    if (ZSTD_isError(args.result)) {
        rb_raise(rb_eRuntimeError, "Decompression failed: %s", ZSTD_getErrorName(args.result));
    }
    rb_str_set_len(result, args.result);
    return result;
}

#format ⇒ Object

#format= ⇒ Object

#initial_capacity ⇒ Object

DCtx initial_capacity getter (instance method)



205
206
207
208
209
210
211
212
213
214
215
 
# File 'ext/vibe_zstd/dctx.c', line 205

static VALUE
vibe_zstd_dctx_get_initial_capacity(VALUE self) {
    vibe_zstd_dctx* dctx;
    TypedData_Get_Struct(self, vibe_zstd_dctx, &vibe_zstd_dctx_type, dctx);

    if (dctx->initial_capacity == 0) {
        // Return the class default
        return vibe_zstd_dctx_get_default_initial_capacity(Qnil);
    }
    return SIZET2NUM(dctx->initial_capacity);
}

#initial_capacity=(value) ⇒ Object

DCtx initial_capacity setter (instance method)



218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
 
# File 'ext/vibe_zstd/dctx.c', line 218

static VALUE
vibe_zstd_dctx_set_initial_capacity(VALUE self, VALUE value) {
    vibe_zstd_dctx* dctx;
    TypedData_Get_Struct(self, vibe_zstd_dctx, &vibe_zstd_dctx_type, dctx);

    if (NIL_P(value)) {
        dctx->initial_capacity = 0;  // Use class default
    } else {
        size_t capacity = NUM2SIZET(value);
        if (capacity == 0) {
            rb_raise(rb_eArgError, "initial_capacity must be positive (or nil to use class default)");
        }
        dctx->initial_capacity = capacity;
    }
    return value;
}

#max_decompressed_size ⇒ Object Also known as: max_size

limit, falling back to the class default. Returns 0 when unlimited.



254
255
256
257
258
259
260
261
262
263
 
# File 'ext/vibe_zstd/dctx.c', line 254

static VALUE
vibe_zstd_dctx_get_max_decompressed_size(VALUE self) {
    vibe_zstd_dctx* dctx;
    TypedData_Get_Struct(self, vibe_zstd_dctx, &vibe_zstd_dctx_type, dctx);

    if (dctx->max_decompressed_size == 0) {
        return SIZET2NUM(default_max_decompressed_size);
    }
    return SIZET2NUM(dctx->max_decompressed_size);
}

#max_decompressed_size=(value) ⇒ Object Also known as: max_size=

DCtx max_decompressed_size setter (instance method); nil = inherit class default



266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
 
# File 'ext/vibe_zstd/dctx.c', line 266

static VALUE
vibe_zstd_dctx_set_max_decompressed_size(VALUE self, VALUE value) {
    vibe_zstd_dctx* dctx;
    TypedData_Get_Struct(self, vibe_zstd_dctx, &vibe_zstd_dctx_type, dctx);

    if (NIL_P(value)) {
        dctx->max_decompressed_size = 0;  // inherit class default
    } else {
        size_t limit = NUM2SIZET(value);
        if (limit == 0) {
            rb_raise(rb_eArgError, "max_decompressed_size must be positive (or nil to inherit the class default)");
        }
        dctx->max_decompressed_size = limit;
    }
    return value;
}

#reset(*args) ⇒ Object

DCtx reset - reset context to clean state



698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
 
# File 'ext/vibe_zstd/dctx.c', line 698

static VALUE
vibe_zstd_dctx_reset(int argc, VALUE* argv, VALUE self) {
    VALUE reset_mode;
    rb_scan_args(argc, argv, "01", &reset_mode);

    vibe_zstd_dctx* dctx;
    TypedData_Get_Struct(self, vibe_zstd_dctx, &vibe_zstd_dctx_type, dctx);

    // Default to SESSION_AND_PARAMETERS if no argument provided
    ZSTD_ResetDirective directive = ZSTD_reset_session_and_parameters;

    if (!NIL_P(reset_mode)) {
        int mode = NUM2INT(reset_mode);
        if (mode == ZSTD_reset_session_only) {
            directive = ZSTD_reset_session_only;
        } else if (mode == ZSTD_reset_parameters) {
            directive = ZSTD_reset_parameters;
        } else if (mode == ZSTD_reset_session_and_parameters) {
            directive = ZSTD_reset_session_and_parameters;
        } else {
            rb_raise(rb_eArgError, "Invalid reset_mode %d: must be ResetDirective::SESSION (1), PARAMETERS (2), or BOTH (3)", mode);
        }
    }

    size_t result = ZSTD_DCtx_reset(dctx->dctx, directive);

    if (ZSTD_isError(result)) {
        rb_raise(rb_eRuntimeError, "Failed to reset decompression context: %s", ZSTD_getErrorName(result));
    }

    return self;
}

#use_prefix(prefix_data) ⇒ Object

DCtx use_prefix - use raw data as prefix (lightweight dictionary)



681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
 
# File 'ext/vibe_zstd/dctx.c', line 681

static VALUE
vibe_zstd_dctx_use_prefix(VALUE self, VALUE prefix_data) {
    vibe_zstd_dctx* dctx;
    TypedData_Get_Struct(self, vibe_zstd_dctx, &vibe_zstd_dctx_type, dctx);

    StringValue(prefix_data);

    size_t result = ZSTD_DCtx_refPrefix(dctx->dctx, RSTRING_PTR(prefix_data), RSTRING_LEN(prefix_data));

    if (ZSTD_isError(result)) {
        rb_raise(rb_eRuntimeError, "Failed to set prefix: %s", ZSTD_getErrorName(result));
    }

    return self;
}

#window_log_max ⇒ Object Also known as: max_window_log

#window_log_max= ⇒ Object Also known as: max_window_log=

DCtx parameter accessors