- `io_read_all` reads all bytes from a file handle into a new buffer. It should only need one allocation per file in cases where the actual file size matches the expected file size. Otherwise it falls back to doubling the buffer size if the actual file size is larger than expected to avoid TOCTOU problems.
- `io_read_all_str` reads all bytes from a file handle into a new buffer and also ensures that the buffer is null-terminated and contains no other null-characters.
- `mem_has_null` is a utility used by `io_read_all_str` to ensure that no null-characters exist in the bytes read from the file.
It didn't have a clear role, it just acted as a distinguisher between
two functions with the same name.
Rename `tw::time_get` to `time_get_nanoseconds` and delete the old
`time_get_nanoseconds`. Move `CCmdlineFix` and the typed
`net_socket_read_wait` function to the global namespace.
This makes the "black console window" less important on Windows (or
anywhere else, for that matter), lets you see logs from other threads in
the f1 console, and removes the distinction between `IConsole::Print`
and `dbg_msg`.
This way new players will get DDNet directory, old ones can switch
directory if they want, or keep using the old one.
If we ever enforce a switch in a future version, this will make it
easier since older DDNet versions will also support the DDNet directory
already.
4818: Add str_startswith_nocase and str_endswith_nocase r=def- a=ChillerDragon
Gets rid of the more complex str_comp_nocase_num usage
## Checklist
- [x] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [x] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [x] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: ChillerDragon <ChillerDragon@gmail.com>
Previously, only Linux used an internal buffer (for optimized
receiving). Now all OSs use an internal buffer so that the call to
`net_udp_recv` behaves the same on all platforms.
- Empty buffers using `io_flush` before calling the OS function for
syncing the file to disk.
- Fix error return of `io_sync` on Windows.
- Don't indiscriminately flush all files on close.
- Add a test that `io_sync` can return without error.
I'm wondering if this helps. Overhead exists, but is not that much (on
my system): 333 ms for initialization instead of 311 ms
If we only want to do this for files written to, then we need to keep
track of how the file was opened.
3638: Revert "Let's not crash the client and server on dbg_assert" r=def- a=heinrich5991
This reverts commit a6e144e.
## Checklist
- [ ] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: heinrich5991 <heinrich5991@gmail.com>
4111: Fix filesystem/IO with unicode filenames on windows r=def- a=Robyt3
It was previously not possible to load a map or really any file with a filename containing unicode, e.g. german umlauts.
Now all filenames are converted from multibyte to widechar on windows, and the correct functions and structs (with the `W` suffix) are used consistently.
Some changes are made to synchronize ddnet's implementation with upstream (once https://github.com/teeworlds/teeworlds/pull/2936 has been merged there):
- The unused and not working `IOFLAG_RANDOM` is removed (on upstream ddnet's `IOFLAG_APPEND` is added instead).
- The `fs_listdir_info` function is renamed to `fs_listdir_fileinfo` and now using the parameter object `CFsFileInfo` that contains the filename, modified date and creation date, making it easier to add more properties in the future. `FS_LISTDIR_INFO_CALLBACK` is renamed to `FS_LISTDIR_CALLBACK_FILEINFO`.
## Checklist
- [X] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: Robert Müller <robert.mueller@uni-siegen.de>
Add tests. Unify behavior of `fs_remove` across operating systems to
only remove files, not directories. Previously on Linux, it would also
delete directories.
3377: Add Thread Safety Analysis r=heinrich5991 a=def-
https://clang.llvm.org/docs/ThreadSafetyAnalysis.html
## Checklist
- [ ] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
3487: Fix centisecs rounding in str_time_float r=heinrich5991 a=def-
<!-- What is the motivation for the changes of this pull request -->
## Checklist
- [ ] Tested the change ingame
- [ ] Provided screenshots if it is a visual change
- [ ] Tested in combination with possibly related configuration options
- [ ] Written a unit test if it works standalone, system.c especially
- [ ] Considered possible null pointers and out of bounds array indexing
- [ ] Changed no physics that affect existing maps
- [ ] Tested the change with [ASan+UBSan or valgrind's memcheck](https://github.com/ddnet/ddnet/#using-addresssanitizer--undefinedbehavioursanitizer-or-valgrinds-memcheck) (optional)
Co-authored-by: def <dennis@felsin9.de>
Purely automatic change. In case of conflict with this change, apply the
other change and rerun the formatting to restore it:
$ python scripts/fix_style.py
Quoting the man page of vsnprintf:
RETURN VALUE
Upon successful return, these functions return the number of characters printed
(excluding the null byte used to end output to strings).
The functions snprintf() and vsnprintf() do not write more than size bytes (including
the terminating null byte ('\0')). If the output was truncated due to this limit,
then the return value is the number of characters (excluding the terminating null
byte) which would have been written to the final string if enough space had been
available. Thus, a return value of size or more means that the output was truncated.
(See also below under NOTES.)
If an output error is encountered, a negative value is returned.
[...]
The glibc implementation of the functions snprintf() and vsnprintf() conforms to the
C99 standard, that is, behaves as described above, since glibc version 2.1. Until
glibc 2.0.6, they would return -1 when the output was truncated.
Edgar
ChillerDragon
