blankhex/bhlib
Archived
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Refactor IO, add buffered IO

Browse Source 
I wasn't happy with existing implementation of the IO, so I decided
to change it - as a result there is no longer BH_IOOpen and BH_IOClose
and many IO operations are now optional (behind BH_IOCtl).

Finnally implemented buffered IO and fixed size memory buffer IO.
This commit is contained in:
Mikhail Romanko
2025-04-26 07:50:13 +03:00
parent 48ddd91dd4
commit 1b6c858a1b
16 changed files with 1473 additions and 1122 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
doc/HowTo/PakReader.md
View File

@@ -61,22 +61,20 @@ To implement this utility, we are going to need to include the following headers
## Working with Files
Working with files in BHLib is based around the IO device (called `BH_IO`).
Firstly, you need to create an IO device with the `BH_FileNew` function.
Secondly, you need to open the IO device with the `BH_IOOpen` function. While
opening the IO device, you can specify in which mode it will work: reading
(`BH_IO_READ`) or writing (`BH_IO_WRITE`). Additionally, we can specify whether
the IO device (or in our case, the file) should exist before opening
(`BH_IO_EXIST`), be truncated before opening (`BH_IO_TRUNCATE`), should it be
created (`BH_IO_CREATE`), or opened in append mode (`BH_IO_APPEND`).
Firstly, you need to create an IO file device with the `BH_FileNew` function.
While doing so, you can specify in which mode it will work: reading
(`BH_FILE_READ`) or writing (`BH_FILE_WRITE`). Additionally, we can specify
whether the file should exist before opening (`BH_IO_EXIST`), be truncated
before opening (`BH_IO_TRUNCATE`), should it be created (`BH_IO_CREATE`), or
opened in append mode (`BH_IO_APPEND`).
Here is an example for opening an existing file in read-only mode:
```c
BH_IO *io = BH_FileNew("coolfile.dat");
if (BH_IOOpen(io, BH_IO_READ | BH_IO_EXIST))
BH_IO *io = BH_FileNew("coolfile.dat", BH_FILE_READ | BH_FILE_EXISTS, NULL);
if (!io)
{
printf("Can't open file 'coolfile.dat'\n", config.file);
BH_IOFree(io);
return -1;
}
```
@@ -207,6 +205,7 @@ Now, let's put everything together and implement `PakReader`.
#define HEADER_SIZE 12
#define ENTRY_SIZE 64
typedef struct PakHeader
{
char id[4];
@@ -361,9 +360,8 @@ static int ProcessPack(Config *config,
if (strcmp(entry.name, config->input))
continue;
output = BH_FileNew(config->output);
if (BH_IOOpen(output, BH_IO_WRITE) ||
BH_IOSeek(io, entry.offset, BH_IO_SEEK_SET) ||
output = BH_FileNew(config->output, BH_FILE_WRITE | BH_FILE_TRUNCATE, NULL);
if (!output || BH_IOSeek(io, entry.offset, BH_IO_SEEK_SET) ||
CopyData(io, output, entry.size))
{
BH_IOFree(output);
@@ -405,8 +403,7 @@ int main(int argc, char **argv)
}
/* Read and write */
io = BH_FileNew(config.file);
if (BH_IOOpen(io, BH_IO_READ | BH_IO_EXIST))
if ((io = BH_FileNew(config.file, BH_FILE_READ | BH_FILE_EXIST, NULL)) == NULL)
{
printf("Can't open file %s\n", config.file);
BH_IOFree(io);

45
doc/HowTo/Utf8Test.md
View File

@@ -21,37 +21,40 @@ To implement this utility, we are going to need to include the following headers
## Working with Files
Working with files in BHLib is based around the IO device (called `BH_IO`).
Firstly, you need to create an IO device with the `BH_FileNew` function.
Secondly, you need to open the IO device with the `BH_IOOpen` function. While
opening the IO device, you can specify in which mode it will work: reading
(`BH_IO_READ`) or writing (`BH_IO_WRITE`). Additionally, we can specify whether
the IO device (or in our case, the file) should exist before opening
(`BH_IO_EXIST`), be truncated before opening (`BH_IO_TRUNCATE`), should it be
created (`BH_IO_CREATE`), or opened in append mode (`BH_IO_APPEND`).
Firstly, you need to create an IO file device with the `BH_FileNew` function.
While doing so, you can specify in which mode it will work: reading
(`BH_FILE_READ`) or writing (`BH_FILE_WRITE`). Additionally, we can specify
whether the file should exist before opening (`BH_IO_EXIST`), be truncated
before opening (`BH_IO_TRUNCATE`), should it be created (`BH_IO_CREATE`), or
opened in append mode (`BH_IO_APPEND`).
Here is an example for opening an existing file in read-only mode:
```c
BH_IO *io = BH_FileNew("coolfile.dat");
if (BH_IOOpen(io, BH_IO_READ | BH_IO_EXIST))
BH_IO *io = BH_FileNew("coolfile.dat", BH_FILE_READ | BH_FILE_EXISTS, NULL);
if (!io)
{
printf("Can't open file 'coolfile.dat'\n", config.file);
BH_IOFree(io);
return -1;
}
```
## Working with UTF-8
Reading UTF-8/UTF-16/UTF-32 is based around simple loop:
1. Read bytes from input (IO or memory) to some buffer.
2. Call `BH_UnicodeDecodeUtf*`. If return value is 0 - we don't have enough data, so go to step 1. Otherwise remove result bytes from the front of the buffer.
3. If readed codepoint equals -1 - we encountered an error, so replace it with the code 0xFFFD.
2. Call `BH_UnicodeDecodeUtf*`. If return value is 0 - we don't have enough
data, so go to step 1. Otherwise remove result bytes from the front of the
buffer.
3. If readed codepoint equals -1 - we encountered an error, so replace it with
the code 0xFFFD.
Writing UTF-8/UTF-16/UTF-32 is straight forward:
1. Call `BH_UnicodeEncodeUtf*`. If return value is 0 - we can't encode codepoint (either codepoint is surrogate pair or outside valid range).
1. Call `BH_UnicodeEncodeUtf*`. If return value is 0 - we can't encode codepoint
(either codepoint is surrogate pair or outside valid range).
2. Write data (to IO or memory).
BH_UnicodeDecodeUtf8(inBuffer, inSize, &unit)
@@ -107,23 +110,25 @@ int main(int argc, char **argv)
if (argc < 2)
printUsage();
inFile = BH_FileNew(argv[1]);
outFile = BH_FileNew(argv[2]);
inFile = BH_FileNew(argv[1], BH_FILE_READ | BH_FILE_EXIST, NULL);
outFile = BH_FileNew(argv[2], BH_FILE_WRITE | BH_FILE_TRUNCATE, NULL);
if (!inFile || BH_IOOpen(inFile, BH_IO_READ | BH_IO_EXIST))
return -1;
if (!outFile || BH_IOOpen(outFile, BH_IO_WRITE | BH_IO_TRUNCATE))
if (!inFile || !outFile)
return -1;
inSize = 0;
while (!(BH_IOFlags(inFile) & BH_IO_FLAG_EOF))
while (1)
{
/* Read one byte and try to decode */
if (!inSize || !(outSize = BH_UnicodeDecodeUtf8(inBuffer, inSize, &unit)))
{
BH_IOPeek(inFile, inBuffer + inSize, 1, &outSize);
BH_IORead(inFile, inBuffer + inSize, 1, &outSize);
inSize += outSize;
if (!outSize)
break;
continue;
}