Verbose CLI + start pulling "Shared-Brotli" (#722)

* Verbose CLI + start pulling "Shared-Brotli"

 * vesbose CLI output; fix #666
 * pull `SHIFT` transforms; currently this is semantically dead code;
   later it will be used by "Shared-Brotli"

2 files changed, 61 insertions, 0 deletions

diff --git a/c/common/transform.c b/c/common/transform.c
index 426e635..c182053 100755
--- a/c/common/transform.c
+++ b/c/common/transform.c
@@ -166,6 +166,7 @@ static BrotliTransforms kBrotliTransforms = {
   kPrefixSuffixMap,
   sizeof(kTransformsData) / (3 * sizeof(kTransformsData[0])),
   kTransformsData,
+  NULL,  /* no extra parameters */
   {0, 12, 27, 23, 42, 63, 56, 48, 59, 64}
 };
 
@@ -190,6 +191,48 @@ static int ToUpperCase(uint8_t* p) {
   return 3;
 }
 
+static int Shift(uint8_t* word, int word_len, uint16_t parameter) {
+  /* Limited sign extension: scalar < (1 << 24). */
+  uint32_t scalar =
+      (parameter & 0x7FFFu) + (0x1000000u - (parameter & 0x8000u));
+  if (word[0] < 0x80) {
+    /* 1-byte rune / 0sssssss / 7 bit scalar (ASCII). */
+    scalar += (uint32_t)word[0];
+    word[0] = (uint8_t)(scalar & 0x7Fu);
+    return 1;
+  } else if (word[0] < 0xC0) {
+    /* Continuation / 10AAAAAA. */
+    return 1;
+  } else if (word[0] < 0xE0) {
+    /* 2-byte rune / 110sssss AAssssss / 11 bit scalar. */
+    if (word_len < 2) return 1;
+    scalar += (uint32_t)((word[1] & 0x3Fu) | ((word[0] & 0x1Fu) << 6u));
+    word[0] = (uint8_t)(0xC0 | ((scalar >> 6u) & 0x1F));
+    word[1] = (uint8_t)((word[1] & 0xC0) | (scalar & 0x3F));
+    return 2;
+  } else if (word[0] < 0xF0) {
+    /* 3-byte rune / 1110ssss AAssssss BBssssss / 16 bit scalar. */
+    if (word_len < 3) return word_len;
+    scalar += (uint32_t)((word[2] & 0x3Fu) | ((word[1] & 0x3Fu) << 6u) |
+        ((word[0] & 0x0Fu) << 12u));
+    word[0] = (uint8_t)(0xE0 | ((scalar >> 12u) & 0x0F));
+    word[1] = (uint8_t)((word[1] & 0xC0) | ((scalar >> 6u) & 0x3F));
+    word[2] = (uint8_t)((word[2] & 0xC0) | (scalar & 0x3F));
+    return 3;
+  } else if (word[0] < 0xF8) {
+    /* 4-byte rune / 11110sss AAssssss BBssssss CCssssss / 21 bit scalar. */
+    if (word_len < 4) return word_len;
+    scalar += (uint32_t)((word[3] & 0x3Fu) | ((word[2] & 0x3Fu) << 6u) |
+        ((word[1] & 0x3Fu) << 12u) | ((word[0] & 0x07u) << 18u));
+    word[0] = (uint8_t)(0xF0 | ((scalar >> 18u) & 0x07));
+    word[1] = (uint8_t)((word[1] & 0xC0) | ((scalar >> 12u) & 0x3F));
+    word[2] = (uint8_t)((word[2] & 0xC0) | ((scalar >> 6u) & 0x3F));
+    word[3] = (uint8_t)((word[3] & 0xC0) | (scalar & 0x3F));
+    return 4;
+  }
+  return 1;
+}
+
 int BrotliTransformDictionaryWord(uint8_t* dst, const uint8_t* word, int len,
     const BrotliTransforms* transforms, int transform_idx) {
   int idx = 0;
@@ -221,6 +264,19 @@ int BrotliTransformDictionaryWord(uint8_t* dst, const uint8_t* word, int len,
         uppercase += step;
         len -= step;
       }
+    } else if (t == BROTLI_TRANSFORM_SHIFT_FIRST) {
+      uint16_t param = (uint16_t)(transforms->params[transform_idx * 2]
+          + (transforms->params[transform_idx * 2 + 1] << 8u));
+      Shift(&dst[idx - len], len, param);
+    } else if (t == BROTLI_TRANSFORM_SHIFT_ALL) {
+      uint16_t param = (uint16_t)(transforms->params[transform_idx * 2]
+          + (transforms->params[transform_idx * 2 + 1] << 8u));
+      uint8_t* shift = &dst[idx - len];
+      while (len > 0) {
+        int step = Shift(shift, len, param);
+        shift += step;
+        len -= step;
+      }
     }
   }
   {
diff --git a/c/common/transform.h b/c/common/transform.h
index 456c12d..b6f86cc 100755
--- a/c/common/transform.h
+++ b/c/common/transform.h
@@ -37,6 +37,8 @@ enum BrotliWordTransformType {
   BROTLI_TRANSFORM_OMIT_FIRST_7 = 18,
   BROTLI_TRANSFORM_OMIT_FIRST_8 = 19,
   BROTLI_TRANSFORM_OMIT_FIRST_9 = 20,
+  BROTLI_TRANSFORM_SHIFT_FIRST = 21,
+  BROTLI_TRANSFORM_SHIFT_ALL = 22,
   BROTLI_NUM_TRANSFORM_TYPES  /* Counts transforms, not a transform itself. */
 };
 
@@ -50,6 +52,9 @@ typedef struct BrotliTransforms {
   uint32_t num_transforms;
   /* Each entry is a [prefix_id, transform, suffix_id] triplet. */
   const uint8_t* transforms;
+  /* Shift for BROTLI_TRANSFORM_SHIFT_FIRST and BROTLI_TRANSFORM_SHIFT_ALL,
+     must be NULL if and only if no such transforms are present. */
+  const uint8_t* params;
   /* Indices of transforms like ["", BROTLI_TRANSFORM_OMIT_LAST_#, ""].
      0-th element corresponds to ["", BROTLI_TRANSFORM_IDENTITY, ""].
      -1, if cut-off transform does not exist. */