| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
if no format scores
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
used (#13135)
* flopdrv: Eliminate floppy_get_drive_type and floppy_set_type
No callers for these functions allowing us to get rid of the
FLOPPY_TYPE_ defines as well.
* flopdrv: Drop flopimg_get_image
* flopdrv: Drop floppy_drive_get_next_id
Only possible user for this is 990_dk under ti99x, but this code is
currently commented out.
* flopdrv: Drop floppy_drive_read_track_data_info_buffer
* flopdrv: Drop floppy_drive_write_track_data_info_buffer
* flopdrv: Drop floppy_drive_format_sector
* flopdrv: Drop floppy_drive_set_index_pulse_callback
Only caller for this is micropolis, but this code is commented
out and callback method does not exist either.
* flopdrv: Drop floppy_drive_get_current_track_size
* flopdrv: Drop floppy_wtd_w
* flopdrv: Drop floppy_wpt_r
This is used in commented out code in agat.
* flopdrv: Drop floppy_dskchg_r
* flopdrv: Drop floppy_twosid_r
* flopdrv: Drop flopimg_get_sectors_per_track and flopimg_get_id_callback
* flopdrv: Drop unneeded state variables: m_id_index, m_rdy
* flopdrv: Make some members as private
|
|
|
|
|
|
|
| |
clang on Windows x86-64 is getting upset about code that is making no
assumptions about the relative sizes of uint64_t and long.
Also, fixed more #include guards.
|
|
|
|
|
|
| |
Pai.
Also fixed a bunch more #include guards.
|
| |
|
|
|
|
| |
stuff with hunks.
|
|
|
|
|
|
|
|
|
| |
- Have metadata_find return std::error_condition instead of throwing an exception
- Replace the is_XXX predicates with check_is_XXX methods that return a std::error_condition, enabling improved error reporting for cdrom_image_device
- Retain read error information in chd_file_compressor
- Make a bunch of methods noexcept
This mostly restores the changes from cc772072fa635146b1df39a5694d2a8f8aa5a34f.
|
| |
|
|
|
|
| |
(#12822)
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
- imagedev/floppy.cpp: Add 3-inch double-sided quad-density format
- oric_microdsk: Add DSQD drives to configuration and enable one by default
- amstrad/pcw.cpp: Upgrade all configurations using double-sided drives to quad density
- bbc_ams3: Make drives single-density since 8271 controller doesn't support MFM
- oric_dsk: Don't blindly trust the number of tracks and sides to be valid
|
| |
|
|
|
|
| |
* bml3mp1805: Use correct SSSD drive type
|
|
|
|
|
| |
* floppy.cpp: Use standard drives for hard-sectored systems
* fix typo
|
|
|
|
|
|
|
| |
Gallager's patches.
Removal of the flux screen experiment.
Removal of the unused get_pos().
|
| |
|
|
|
|
|
|
|
|
|
|
| |
* sony_news.xml: Added proper compatibility flags for different
generations.
* cpu/pic16x8x: This is very much a derivative work.
* cpu/tms32025.cpp: Allow stack push/pop to be inlined.
* tecmo/bombjack.cpp: Avoid needing to remove and replace devices in
machine configuration.
* Various other cleanup.
|
|
|
|
| |
disk without image file.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently, when `floppy_sound_device::step` selects a new seek sample, the
value of `m_seek_samplepos` remains the same as it was the last time a seek
sample was played. This might be an invalid position for the new sample,
since the seek samples have different lengths. When this is the case, the next
call to `floppy_sound_device::sound_stream_update` makes an out-of-bounds memory
access as it reads the sample data.
Fix this by resetting `m_seek_samplepos` to 0 when a new seek sample is
selected.
This also makes the seek sounds more consistent, since previously they would
start from a semi-random position each time.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
`floppy_image_device::find_index` uses binary search to find the index for
which `buf[spos] <= position < buf[spos + 1]`. However, the algorithm behaves
incorrectly when `position < buf[0]`. In this case, the algorithm returns 0,
as if `position` was between `buf[0]` and `buf[1]`.
The effect of this is that if `get_next_transition` is called with a timestamp
that is between the start of the revolution and the first transition, then
instead of returning the timestamp of that transition, it returns the timestamp
of the second transition instead. Essentially, the first 1-bit on the track gets
flipped to a 0.
I have encountered this in Apple II emulation, where this bug manifests as
sporadic I/O errors.
Fix it by doing two things:
1. Replace `find_index` with a call to `upper_bound` from the standard library,
which behaves correctly in edge cases.
2. If `upper_bound` signals that `position < buf[0]`, then adjust `base` and
`index` to point to the last transition of the previous revolution.
|
|
|
|
|
|
|
|
|
|
|
| |
CD-ROMs, indexes, and track flags (#12201)
* util/cdrom: Refactoring
* util/cdrom: Read all indexes from cue files
* util/cdrom: Read in track flags from cues
* util/cdrom: Multisession support for cues
* machine/t10mmc: Playback from current head, other misc cleanup
* machine/t10mmc: Implement T10MMC_CMD_MECHANISM_STATUS
* machine/t10mmc: Implement TOC_FORMAT_FULL_TOC, TOC_FORMAT_SESSIONS, and return proper indexes for T10MMC_CMD_READ_SUB_CHANNEL
|
|
|
|
|
| |
cassette images in FLAC format. (#12115)
util/flac.cpp: Implemented seek/tell callbacks for FLAC library.
|
| |
|
|
|
|
|
| |
emu/diimage.h: Removed fread overloads that allocate memory for output.
util/core_file.cpp: Changed output size of load to size_t.
|
|
|
|
| |
* util/options.h: Use forwarding header
|
|
|
|
|
| |
-imagedev/cdrom.cpp: Report unsuitable CHDs as "invalid image" rather
than "unsupported operation".
|
|
|
|
|
|
|
| |
ROM region.
* Reverted workaround 4c0957d7f0ba6a72e7267270db07ee06fe717109.
* Cleaned up a few things.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
| |
------------------------------
Cuty Line Limited (ver.1.30) [freaktech]
- igs/goldstar.cpp: enabled girls display for cmast91 and other games on same hw, removed 2nd ppi as PCB pics and readme only show one
- various files: removed superfluous semicolons
|
|
|
|
|
|
|
|
|
| |
This caused numerous regressions, including floppy formats not working
and application hangs when using some formats.
This reverts commit b2c399c61d65063ae95e8387d34b098e9516b1a9.
This reverts commit cf510adc9d34b4ad9a065f3d5bfbcf17115ea06f.
|
|
|
|
| |
This reverts commit cc772072fa635146b1df39a5694d2a8f8aa5a34f.
|
|
|
|
|
|
|
| |
- Change is_XXX to return std::error_condition instead of bool so that errors can be passed down the line; rename these to check_is_XXX to indicate that they are no longer basic predicates
- Change return type of internal function metadata_find to std::error_condition so that errors can be returned rather than thrown
* imagedev/cdromimg.cpp: Fix bug where cdrom_file object could be constructed twice in a row for CD-ROMs and once for DVDs
|
| |
|
| |
|
|
|
|
|
| |
- Add noexcept to format information overrides and several simple floppy_image accessors.
- Be more const-correct and use references instead of raw pointers wherever applicable. In particular, floppy_image_device::load and floppy_image_device::save now take a reference to floppy_image rather than a pointer, const-qualified in the latter case.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
render/drawbgfx.cpp: Return an error if Wayland EGL surface can't be
created for additional windows.
emu/emucore.h: Added explicitly defaulted copy and move constructors for
emu_fatalerror.
Fixed apparent misunderstanding of const. Returning const value types
and casting to const value types is pointless outside very narrow use
cases. Putting const value type parameters in interfaces just makes
trouble.
cpu/adsp2100: Use count_leading_ones_32 where it's simple rather than
inverting and counting leading zeroes.
util/multibyte.h: Don't pollute global namespace, constexpr implies
inline, make narrowing casts explicit.
imagedev/simh_tape_image.h: inline is implied for member functions with
bodies supplied at declaration.
Tidied up some ugly casts in various places.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Read callbacks now need a default return value supplied at construction.
Replaced isnull() with isunset() which tells you if the callback wasn't
configured rather than whether it isn't safe to call.
Enabled validation of device callbacks (it seems it was disabled at some
point, probably accidentally).
Device callbacks and object finders now implement the same interface for
resolution.
|
| |
|
| |
|
|
|
|
| |
Belmont]
|
|
|
|
| |
function signatures. (#11283) [Ryan Holtz]
|
| |
|