| Rev |
Age |
Author |
Path |
Log message |
Diff |
| 9935 |
251 d 22 h |
spaceraven |
/programs/demos/ScreenRuler/ |
Deleting binary; correcting small mistakes in ScreenRuler.asm |
|
| 9934 |
251 d 23 h |
spaceraven |
/programs/demos/ScreenRuler/ |
New program - Screen Ruler, for measuring distance between 2 pixel on screen. |
|
| 9933 |
261 d 15 h |
dunkaist |
/programs/system/taskbar/trunk/ |
[taskbar] Add LCtrl+Esc hotkey to open/close menu |
|
| 9932 |
273 d 19 h |
Doczom |
/kernel/trunk/ |
[KERNEL] cleared the code of unnecessary comments |
|
| 9931 |
274 d 6 h |
boppan |
/kernel/trunk/ |
[KERNEL][TEST] Run UMKa tests if enabled.
Before this commit UMKa was downloaded, but none of its tests had been executed. Now it executes all the tagged UMKa tests (41 for now). Some tags had been excluded, like #acpi, because the ACPI subsystem is still under development. |
|
| 9930 |
275 d 20 h |
Doczom |
/kernel/trunk/ |
[KERNEL] The window subsystem has been redesigned:
-extended WDATA structures up to 128 bytes
-added the background_windows constant
-the draw_data array has been deleted
-APPDATA.window and WDATA.thread fields added
-minor fixes in the operation of external file system drivers |
|
| 9929 |
279 d 22 h |
boppan |
/kernel/trunk/ |
[KERNEL] Trigger GitHub mirror sync. |
|
| 9928 |
287 d 21 h |
boppan |
/_tools/workspace/ |
[WS] Make run.py work on modern qemu |
|
| 9927 |
288 d 2 h |
boppan |
/programs/develop/clink/ |
[clink] Make it able to read and write more than 64k - 1 COFF relocation
These are changes from two commits from the upstream clink repo (https://github.com/mkostoevr/clink) and one commit from the upstream epep repo (https://github.com/mkostoevr/epep).
clink:
Commit e63ed12ead17e47d77e848da0e7b9f4dd3ad5127
Bugfix: Make it able to write more than 64k - 1 COFF relocations
If the relocation count is greater than 0xffff then it can't fit
in the NumberOfRelocations field of a section header. In order to
specify greater relocation count IMAGE_SCN_LNK_NRELOC_OVFL flag
should be added to Characteristics and the actual relocation count
should be written into VirtualAddress field of the first COFF
relocation.
Commit 328fc3112a30fcaa808f2cad34028a6507d602a3
Bugfix: Make it able to read more than 64k - 1 COFF relocations
New EPEP API allows to read more than 64k - 1 relocations. Use it
when reading relocation count for filling SectionInfo and when
reading relocations for wriring the output file.
epep:
Commit 3bed4e348a486c346d0a452c58c1d85e1805c09d
Feature: Higher-level COFF relocations API
Number of COFF relocations of a section is stored in the 16-bit
NumberOfRelocations field of a section header. If a COFF object has
more than 2^16 - 1 relocations, then the value does not fit in the
field.
To solve this problem, IMAGE_SCN_LNK_NRELOC_OVFL flag of a section
header has been introduced. If this flag is set for the section,
then the actual number of relocations is stored in the
VirtualAddress field of the first relocation.
If the flag is set, then NumberOfRelocations field of the section
header should be equal to 0xffff, othervice the linker should give
an error.
So this patch introduces few functions adressing this mechanism.
epep_section_contains_extended_relocations:
Checks whether the section has more than 2^16 - 1 relocations.
Retrns error if the IMAGE_SCN_LNK_NRELOC_OVFL flag is set, but
the NumberOfRelocations field is not equal to 0xffff.
epep_get_section_extended_number_of_relocations:
Reads the number of COFF relocations from the VirtualAddress
field of the first COFF relocation.
epep_get_section_number_of_relocations_x:
Gives the number of meaningful relocations of the section.
If the section has less than 2^16 relocations, then returns the
value of the NumberOfRelocations field of the section header,
othervice reads the number of relocations from the first COFF
relocation, but: since the first relocation in this case is not
meaningful, gives the actual number of relocations minus one.
This is used in the function documented below.
Returns 1 in the last argument if the section contains extended
number of relocations, 0 othervice.
epep_get_section_relocation_by_index_x:
If the section has less than 2^16 relocations, then just reads
a relocation by the given index. In case if the section has
extended number of relocations, the first relocation is not
meaningful, so it is skipped, and the relocation at index + 1
is read instead. |
|
| 9926 |
330 d 0 h |
Doczom |
/kernel/trunk/ |
[KERNEL] the window subsystem was refactored, WDATA structures were expanded and magic numbers were replaced with constants |
|
| 9925 |
333 d 13 h |
Doczom |
/kernel/trunk/ |
[KERNEL] fixed a bug in rev 9917 |
|
| 9924 |
335 d 1 h |
Doczom |
/ |
[SDHCI] Added new driver on SD host controller |
|
| 9923 |
336 d 22 h |
Boppan |
/kernel/trunk/ |
[KERNEL][TEST] Download and build UMKa if enabled. |
|
| 9922 |
337 d 2 h |
Boppan |
/kernel/trunk/ |
[KERNEL][TEST] Initialize tup if it is not. |
|
| 9921 |
337 d 3 h |
Boppan |
/kernel/trunk/test/common/ |
[KERNEL][TEST] Fix tests fail for new qemu. |
|
| 9920 |
337 d 3 h |
Boppan |
/kernel/trunk/ |
[KERNEL][TEST] Print informative message on test fail. |
|
| 9919 |
337 d 4 h |
Boppan |
/kernel/trunk/test/common/ |
[KERNEL][TEST] Fail on prematurely finished qemu. |
|
| 9918 |
337 d 21 h |
Boppan |
/kernel/trunk/ |
Tests: Fix testing kernel build
The test system used a custom command to build the testing kernel.
Make it use the tup system. |
|
| 9917 |
337 d 22 h |
Doczom |
/kernel/trunk/ |
[KERNEL]: Replaced magic numbers with constants and cleaned code |
|
| 9916 |
347 d 11 h |
Doczom |
/programs/network/ftpd/ |
FTPd: added the DELE and STOR functions, fixed the output code for the new connection |
|
| 9915 |
378 d 23 h |
akron1 |
/programs/develop/cedit/ |
CEdit: added [shift+mouse wheel] for horizontal scrolling;
search improved. |
|
| 9914 |
385 d 5 h |
qullarwee |
/programs/other/mousepos/ |
Upload "mousepos" utility |
|
| 9913 |
402 d 21 h |
akron1 |
/programs/develop/cedit/ |
CEdit: bugfix |
|
| 9912 |
404 d 19 h |
akron1 |
/ |
FB2 Reader: fixed rolling up |
|
| 9911 |
406 d 22 h |
Doczom |
/kernel/trunk/ |
[KERNEL] cleaned code and added description of some functions |
|
| 9910 |
407 d 19 h |
Doczom |
/kernel/trunk/ |
[KERNEL] updated local labels in functions and other small changes |
|
| 9909 |
418 d 18 h |
akron1 |
/programs/develop/cedit/ |
CEdit:
- highlight @@:/@f/@b/@r
- numpad support |
|
| 9908 |
419 d 0 h |
Doczom |
/programs/ |
Updated dll.load. Added support for loading the library using the full path.
Fixed makefile for Whether |
|
| 9907 |
429 d 4 h |
akron1 |
/programs/develop/cedit/ |
CEdit: fixed whole word search |
|
| 9906 |
441 d 19 h |
akron1 |
/programs/develop/cedit/ |
CEdit: minor fix |
|
| 9905 |
442 d 23 h |
leency |
/data/common/File Managers/ndn/ |
NDN: upload a new version v3.00.0008 2022-12-26 from http://ndn.muxe.com/download/ |
|
| 9904 |
448 d 21 h |
akron1 |
/programs/develop/cedit/ |
CEdit: optimization |
|
| 9903 |
451 d 2 h |
akron1 |
/programs/develop/cedit/ |
CEdit: reduced memory usage, small improvements |
|
| 9902 |
453 d 1 h |
akron1 |
/programs/develop/cedit/ |
CEdit: reduced memory usage |
|
| 9901 |
453 d 1 h |
akron1 |
/data/ |
move fNav to ISO |
|
| 9900 |
460 d 17 h |
Doczom |
/kernel/trunk/core/ |
[KERNEL]: Replaced magic numbers with constants:
PAGE_SIZE, -PAGE_SIZE, PAGE_SIZE-1 |
|
| 9899 |
462 d 19 h |
dunkaist |
/kernel/trunk/ |
Unhardcode some PAGE_SIZE related numbers
4096 is PAGE_SIZE
not 4095 is -PAGE_SIZE
etc |
|
| 9898 |
474 d 3 h |
akron1 |
/ |
FB2 Reader: refactoring, reduced memory usage for images, small improvements |
|
| 9897 |
476 d 13 h |
dunkaist |
/kernel/trunk/ |
[kernel] Fix reading from empty board buffer (sf63.2)
Syscall 63.2, reading from the board buffer, was overwriting registers
edx and ebp when the buffer was empty. Now it sets eax and ebx according
to the kernel API.
Because of this ancient bug BOARD checks the status of sf63.2 syscall
this way:
mcall 63, 2
cmp ebx, 1
jne no_data
; when the buffer is empty, ebx is untouched, i.e. still 2, haha
; edx and ebp are destroyed, the code is lucky not to use them
By the way, the bug was found using umka tool. |
|
| 9896 |
484 d 2 h |
akron1 |
/ |
FB2 Reader: upload source, small changes |
|