database/sqlite
1
0
Fork 0
mirror of https://github.com/sqlite/sqlite.git synced 2025-08-07 02:42:48 +03:00
Code Activity

Add sqlite3_wasm_vfs_create_file() to replace Emscripten's FS.createDataFile() in a (mostly) VFS-agnostic way. Add a test for worker1's export (to bytearray) support. Re-add worker1 open-from-bytearray using sqlite3_wasm_vfs_create_file() but it's untested (requires a new interactive test app or maybe reconsideration).

Browse Source 
FossilOrigin-Name: b35e1225c91a3cadc0d25af1e4e790237256d194990faa13190e343ed03e11c5
This commit is contained in:
stephan
2022-11-02 11:53:31 +00:00
parent faff0410dc
commit f45c33701d
10 changed files with 264 additions and 66 deletions
Download Patch File Download Diff File Expand all files Collapse all files

102
ext/wasm/api/sqlite3-wasm.c
View File

@@ -12,7 +12,6 @@
**
** emcc -o sqlite3.wasm ... -I/path/to/sqlite3-c-and-h sqlite3-wasm.c
*/
#define SQLITE_WASM
#ifdef SQLITE_WASM_ENABLE_C_TESTS
/*
@@ -928,8 +927,8 @@ int sqlite3_wasm_db_export_chunked( sqlite3* pDb,
** sqlite3_free() to free it.
*/
SQLITE_WASM_KEEP
int sqlite3_wasm_db_serialize( sqlite3* pDb, unsigned char **pOut,
sqlite3_int64 * nOut, unsigned int mFlags ){
int sqlite3_wasm_db_serialize( sqlite3 *pDb, unsigned char **pOut,
sqlite3_int64 *nOut, unsigned int mFlags ){
unsigned char * z;
if( !pDb || !pOut ) return SQLITE_MISUSE;
if(nOut) *nOut = 0;
@@ -942,6 +941,103 @@ int sqlite3_wasm_db_serialize( sqlite3* pDb, unsigned char **pOut,
}
}
/*
** This function is NOT part of the sqlite3 public API. It is strictly
** for use by the sqlite project's own JS/WASM bindings.
**
** Creates a new file using the I/O API of the given VFS, containing
** the given number of bytes of the given data. If the file exists,
** it is truncated to the given length and populated with the given
** data.
**
** This function exists so that we can implement the equivalent of
** Emscripten's FS.createDataFile() in a VFS-agnostic way. This
** functionality is intended for use in uploading database files.
**
** If pVfs is NULL, sqlite3_vfs_find(0) is used.
**
** If zFile is NULL, pVfs is NULL (and sqlite3_vfs_find(0) returns
** NULL), or nData is negative, SQLITE_MISUSE are returned.
**
** On success, it creates a new file with the given name, populated
** with the fist nData bytes of pData. If pData is NULL, the file is
** created and/or truncated to nData bytes.
**
** Whether or not directory components of zFilename are created
** automatically or not is unspecified: that detail is left to the
** VFS. The "opfs" VFS, for example, create them.
**
** Not all VFSes support this functionality, e.g. the "kvvfs" does
** not.
**
** If an error happens while populating or truncating the file, the
** target file will be deleted (if needed) if this function created
** it. If this function did not create it, it is not deleted but may
** be left in an undefined state.
**
** Returns 0 on success. On error, it returns a code described above
** or propagates a code from one of the I/O methods.
**
** Design note: nData is an integer, instead of int64, for WASM
** portability, so that the API can still work in builds where BigInt
** support is disabled or unavailable.
*/
SQLITE_WASM_KEEP
int sqlite3_wasm_vfs_create_file( sqlite3_vfs *pVfs,
const char *zFilename,
const unsigned char * pData,
int nData ){
int rc;
sqlite3_file *pFile = 0;
sqlite3_io_methods const *pIo;
const int openFlags = SQLITE_OPEN_READWRITE | SQLITE_OPEN_CREATE;
int flagsOut = 0;
int fileExisted = 0;
int doUnlock = 0;
const unsigned char *pPos = pData;
const int blockSize = 512
/* Because we are using pFile->pMethods->xWrite() for writing, and
** it may have a buffer limit related to sqlite3's pager size, we
** conservatively write in 512-byte blocks (smallest page
** size). */;
if( !pVfs ) pVfs = sqlite3_vfs_find(0);
if( !pVfs || !zFilename || nData<0 ) return SQLITE_MISUSE;
pVfs->xAccess(pVfs, zFilename, SQLITE_ACCESS_EXISTS, &fileExisted);
rc = sqlite3OsOpenMalloc(pVfs, zFilename, &pFile, openFlags, &flagsOut);
if(rc) return rc;
pIo = pFile->pMethods;
if( pIo->xLock ) {
/* We need xLock() in order to accommodate the OPFS VFS, as it
** obtains a writeable handle via the lock operation and releases
** it in xUnlock(). If we don't do those here, we have to add code
** to the VFS to account check whether it was locked before
** xFileSize(), xTruncate(), and the like, and release the lock
** only if it was unlocked when the op was started. */
rc = pIo->xLock(pFile, SQLITE_LOCK_EXCLUSIVE);
doUnlock = 0==rc;
}
if( 0==rc) rc = pIo->xTruncate(pFile, nData);
if( 0==rc && 0!=pData && nData>0 ){
while( 0==rc && nData>0 ){
const int n = nData>=blockSize ? blockSize : nData;
rc = pIo->xWrite(pFile, pPos, n, (sqlite3_int64)(pPos - pData));
nData -= n;
pPos += n;
}
if( 0==rc && nData>0 ){
assert(nData<512);
rc = pIo->xWrite(pFile, pPos, nData, (sqlite3_int64)(pPos - pData));
}
}
if( pIo->xUnlock && doUnlock!=0 ) pIo->xUnlock(pFile, SQLITE_LOCK_NONE);
pIo->xClose(pFile);
if( rc!=0 && 0==fileExisted ){
pVfs->xDelete(pVfs, zFilename, 1);
}
return rc;
}
/*
** This function is NOT part of the sqlite3 public API. It is strictly
** for use by the sqlite project's own JS/WASM bindings.