diff --git a/manifest b/manifest index 1324137007..bf26b0611c 100644 --- a/manifest +++ b/manifest @@ -1,5 +1,5 @@ -C Removed\snever-used\ssymbol\sSQLITE_ResultDetails.\s(CVS\s965) -D 2003-05-10T03:04:34 +C Changes\sto\scomments.\s\sIn\sparticular,\sa\slengthy\scomment\swas\sadded\sto\sencode.c\r\nthat\sexplains\show\sthe\sencoder\salgorithm\sworks.\s(CVS\s966) +D 2003-05-10T03:36:54 F Makefile.in 004acec253ecdde985c8ecd5b7c9accdb210378f F Makefile.linux-gcc b86a99c493a5bfb402d1d9178dcdc4bd4b32f906 F README f1de682fbbd94899d50aca13d387d1b3fd3be2dd @@ -20,14 +20,14 @@ F spec.template 238f7db425a78dc1bb7682e56e3834c7270a3f5e F sqlite.1 83f4a9d37bdf2b7ef079a82d54eaf2e3509ee6ea F sqlite.pc.in 30552343140c53304c2a658c080fbe810cd09ca2 F src/attach.c 7ebc7487de43e357a64226f8abef81f2669f2183 -F src/auth.c 53b8923f17f364af84501fa99dc00c779913a26d +F src/auth.c c8f50d4507e37779d96ff3c55417bc2b612dfed6 F src/btree.c 077d75aee4ed63f3628698611ba43c87097d458d F src/btree.h 23c50e00709de16a5dce1fcea9c0429cc955ff0e F src/btree_rb.c 8e00e40be264716e1929987878672e55d9e0e76d F src/build.c e24461d42381a36de88de6af06c03d9f14588705 F src/copy.c 44b13fd4d2444fb53bff8a5ecee1c5f6f86a8263 F src/delete.c f9536a75b444a21f11b7a1bc0fb8c876f691b013 -F src/encode.c faf03741efe921755ec371cf4a6984536de00042 +F src/encode.c ed720e54ec4ef4d4de651592f1dd1c74d422bbd2 F src/expr.c a666ef5220ca90ebcf40c8ccc783966a345a7616 F src/func.c 882c3ed5a02be18cd904715c7ec62947a34a3605 F src/hash.c 4fc39feb7b7711f6495ee9f2159559bedb043e1f @@ -59,7 +59,7 @@ F src/trigger.c 8ee811986080de60d9d883ad96daffea82014f27 F src/update.c dc3b680b4cf2cb6d839950b68d632850746639b9 F src/util.c 87635cfdfffa056a8d3147719357aa442374f78c F src/vacuum.c 14ac3073203fa021e01ffe33db56968ad79a8344 -F src/vdbe.c 829fa31c2169588db7a041950cb95c6bb75811a6 +F src/vdbe.c e9e560b0c568fb4ffb189d3e0d91628a33d2a367 F src/vdbe.h 985c24f312d10f9ef8f9a8b8ea62fcdf68e82f21 F src/where.c ef11773a07cf075740378d7d1cbabf328146fdc9 F test/all.test 569a92a8ee88f5300c057cc4a8f50fbbc69a3242 @@ -165,7 +165,7 @@ F www/speed.tcl cb4c10a722614aea76d2c51f32ee43400d5951be F www/sqlite.tcl 4bd1729e320f5fa9125f0022b281fbe839192125 F www/tclsqlite.tcl 1db15abeb446aad0caf0b95b8b9579720e4ea331 F www/vdbe.tcl 2013852c27a02a091d39a766bc87cff329f21218 -P bf558e79524c04339f1174dc935e39d47cb33457 -R dbcdc1e6ba937154b079e14cdd4eab7b -U jplyon -Z c717a1a241516bca947b8579016988c0 +P 26bd7bb1779e14f19c00e450aaa9529f2aabf131 +R b680c0a504530ce6590f8a3af4ddd816 +U drh +Z d87b54e38f0dd3d4af38249efc6d41d2 diff --git a/manifest.uuid b/manifest.uuid index b0732489e7..5f5d23af07 100644 --- a/manifest.uuid +++ b/manifest.uuid @@ -1 +1 @@ -26bd7bb1779e14f19c00e450aaa9529f2aabf131 \ No newline at end of file +8b388b2f690dbfd50eefc0fdede1c7785f78afa1 \ No newline at end of file diff --git a/src/auth.c b/src/auth.c index 4bd4903f9e..a7566d7487 100644 --- a/src/auth.c +++ b/src/auth.c @@ -14,7 +14,7 @@ ** systems that do not need this facility may omit it by recompiling ** the library with -DSQLITE_OMIT_AUTHORIZATION=1 ** -** $Id: auth.c,v 1.9 2003/05/02 14:32:13 drh Exp $ +** $Id: auth.c,v 1.10 2003/05/10 03:36:54 drh Exp $ */ #include "sqliteInt.h" @@ -28,15 +28,35 @@ ** Set or clear the access authorization function. ** ** The access authorization function is be called during the compilation -** phase to verify that the user has read and/or write access permission +** phase to verify that the user has read and/or write access permission on ** various fields of the database. The first argument to the auth function ** is a copy of the 3rd argument to this routine. The second argument ** to the auth function is one of these constants: ** -** SQLITE_READ_COLUMN -** SQLITE_WRITE_COLUMN -** SQLITE_DELETE_ROW -** SQLITE_INSERT_ROW +** SQLITE_COPY +** SQLITE_CREATE_INDEX +** SQLITE_CREATE_TABLE +** SQLITE_CREATE_TEMP_INDEX +** SQLITE_CREATE_TEMP_TABLE +** SQLITE_CREATE_TEMP_TRIGGER +** SQLITE_CREATE_TEMP_VIEW +** SQLITE_CREATE_TRIGGER +** SQLITE_CREATE_VIEW +** SQLITE_DELETE +** SQLITE_DROP_INDEX +** SQLITE_DROP_TABLE +** SQLITE_DROP_TEMP_INDEX +** SQLITE_DROP_TEMP_TABLE +** SQLITE_DROP_TEMP_TRIGGER +** SQLITE_DROP_TEMP_VIEW +** SQLITE_DROP_TRIGGER +** SQLITE_DROP_VIEW +** SQLITE_INSERT +** SQLITE_PRAGMA +** SQLITE_READ +** SQLITE_SELECT +** SQLITE_TRANSACTION +** SQLITE_UPDATE ** ** The third and fourth arguments to the auth function are the name of ** the table and the column that are being accessed. The auth function diff --git a/src/encode.c b/src/encode.c index a4bea5574c..b313d4f821 100644 --- a/src/encode.c +++ b/src/encode.c @@ -15,10 +15,90 @@ ** data in an SQLite database. The code in this file is not used by any other ** part of the SQLite library. ** -** $Id: encode.c,v 1.5 2003/01/19 03:59:46 drh Exp $ +** $Id: encode.c,v 1.6 2003/05/10 03:36:54 drh Exp $ */ #include +/* +** How This Encoder Works +** +** The output is allowed to contain any character except 0x27 (') and +** 0x00. This is accomplished by using an escape character to encode +** 0x27 and 0x00 as a two-byte sequence. The escape character is always +** 0x01. An 0x00 is encoded as the two byte sequence 0x01 0x01. The +** 0x27 character is encoded as the two byte sequence 0x01 0x03. Finally, +** the escape character itself is encoded as the two-character sequence +** 0x01 0x02. +** +** To summarize, the encoder works by using an escape sequences as follows: +** +** 0x00 -> 0x01 0x01 +** 0x01 -> 0x01 0x02 +** 0x27 -> 0x01 0x03 +** +** If that were all the encoder did, it would work, but in certain cases +** it could double the size of the encoded string. For example, to +** encode a string of 100 0x27 character would require 100 instances of +** the 0x01 0x03 escape sequence resulting in a 200-character output. +** We would prefer to keep the size of the encoded string smaller than +** this. +** +** To minimize the encoding size, we first add a fixed offset value to each +** byte in the sequence. The addition is module 256. (That is to say, if +** the sum of the original character value and the offset exceeds 256, then +** the higher order bits are truncated.) The offset is chosen to minimize +** the number of characters in the string that need to be escaped. For +** example, in the case above where the string was composed of 100 0x27 +** characters, the offset might be 0x01. Each of the 0x27 characters would +** then be converted into an 0x28 character which would not need to be +** escaped at all and so the 100 character input string would be converted +** into just 100 characters of output. Actually 101 characters of output - +** we have to record the offset used as the first byte in the sequence so +** that the string can be decoded. Since the offset value is stored as +** part of the output string and the output string is not allowed to contain +** characters 0x00 or 0x27, the offset cannot be 0x00 or 0x27. +** +** Here, then, are the encoding steps: +** +** (1) Choose an offset value and make it the first character of +** output. +** +** (2) Copy each input character into the output buffer, one by +** one, adding the offset value as you copy. +** +** (3) If the value of an input character plus offset is 0x00, replace +** that one character by the two-character sequence 0x01 0x01. +** If the sum is 0x01, replace it with 0x01 0x02. If the sum +** is 0x27, replace it with 0x01 0x03. +** +** (4) Put a 0x00 terminator at the end of the output. +** +** Decoding is obvious: +** +** (5) Copy encoded characters except the first into the decode +** buffer. Set the first encoded character aside for use as +** the offset in step 7 below. +** +** (6) Convert each 0x01 0x01 sequence into a single character 0x00. +** Convert 0x01 0x02 into 0x01. Convert 0x01 0x03 into 0x27. +** +** (7) Subtract the offset value that was the first character of +** the encoded buffer from all characters in the output buffer. +** +** The only tricky part is step (1) - how to compute an offset value to +** minimize the size of the output buffer. This is accomplished to testing +** all offset values and picking the one that results in the fewest number +** of escapes. To do that, we first scan the entire input and count the +** number of occurances of each character value in the input. Suppose +** the number of 0x00 characters is N(0), the number of occurances of 0x01 +** is N(1), and so forth up to the number of occurances of 0xff is N(256). +** An offset of 0 is not allowed so we don't have to test it. The number +** of escapes required for an offset of 1 is N(1)+N(2)+N(40). The number +** of escapes required for an offset of 2 is N(2)+N(3)+N(41). And so forth. +** In this way we find the offset that gives the minimum number of escapes, +** and thus minimizes the length of the output string. +*/ + /* ** Encode a binary buffer "in" of size n bytes so that it contains ** no instances of characters '\'' or '\000'. The output is diff --git a/src/vdbe.c b/src/vdbe.c index 511e76cb05..c67bd423ea 100644 --- a/src/vdbe.c +++ b/src/vdbe.c @@ -36,7 +36,7 @@ ** in this file for details. If in doubt, do not deviate from existing ** commenting and indentation practices when changing or adding code. ** -** $Id: vdbe.c,v 1.221 2003/05/02 14:32:14 drh Exp $ +** $Id: vdbe.c,v 1.222 2003/05/10 03:36:54 drh Exp $ */ #include "sqliteInt.h" #include @@ -3453,10 +3453,10 @@ case OP_OpenWrite: { /* Opcode: OpenTemp P1 P2 * ** -** Open a new cursor that points to a table or index in a temporary -** database file. The temporary file is opened read/write even if -** the main database is read-only. The temporary file is deleted -** when the cursor is closed. +** Open a new cursor to a transient table. +** The transient cursor is always opened read/write even if +** the main database is read-only. The transient table is deleted +** automatically when the cursor is closed. ** ** The cursor points to a BTree table if P2==0 and to a BTree index ** if P2==1. A BTree table must have an integer key and can have arbitrary @@ -3464,7 +3464,7 @@ case OP_OpenWrite: { ** ** This opcode is used for tables that exist for the duration of a single ** SQL statement only. Tables created using CREATE TEMPORARY TABLE -** are opened using OP_OpenAux or OP_OpenWrAux. "Temporary" in the +** are opened using OP_OpenRead or OP_OpenWrite. "Temporary" in the ** context of this opcode means for the duration of a single SQL statement ** whereas "Temporary" in the context of CREATE TABLE means for the duration ** of the connection to the database. Same word; different meanings. @@ -3997,8 +3997,8 @@ case OP_Delete: { /* Opcode: KeyAsData P1 P2 * ** ** Turn the key-as-data mode for cursor P1 either on (if P2==1) or -** off (if P2==0). In key-as-data mode, the Field opcode pulls -** data off of the key rather than the data. This is useful for +** off (if P2==0). In key-as-data mode, the OP_Column opcode pulls +** data off of the key rather than the data. This is used for ** processing compound selects. */ case OP_KeyAsData: {