2020 by C. Masloch. Usage of the works is permitted provided that this instrument is retained with the works, so that any entity that uses the works is notified of this instrument. DISCLAIMER: THE WORKS ARE WITHOUT WARRANTY.
This document has been compiled on 2021-02-15.
lDebug is a 86-DOS debugger based on the MS-DOS Debug clone FreeDOS Debug. It features DPMI client support for 32-bit and 16-bit segments, a 686-level assembler and disassembler, an expression evaluator, an InDOS and a bootloaded mode, script file reading, serial port I/O, permanent breakpoints, conditional tracing, buffered tracing, and auto-repetition of some commands. There is also a symbolic debugging branch being developed.
Building lDebug is not supported on conventional DOS-like systems. (DJGPP environments may suffice but are not tested.)
The following components are required to build with the provided scripts:
%deftok’ is not supported
%deftok’ is implemented wrongly
%assign %$foo%[bar] quux’ doesn't function right
hg pull’ to update the repo
hg up default’ or the symbolic branch with ‘hg up symbolic’ or any other available commit you want to build
hg fetch’ or the sequence of ‘hg pull’ then ‘hg up’ to update the repos. (Usually the additional source repos do not have multiple branches.)
INICOMP_METHOD in ovr.sh to select none, one, or several compression methods. Surround multiple values with quotes and delimit with blanks. If the value "none" is used no compression will occur. If several values are given the smallest of the resulting files will be used as the ldebug.com result. This favours LZMA-lzip (lzd) and Exomizer 3 (exodecr) compression as they result in the best ratios. The uncompressed ldebugu.com file will always be generated, you can rename or copy or symlink it to use it as ldebug.com if you want.
use_build_decomp_test option. This insures that the compressed executables will actually succeed in decompression when entered in EXE mode, and will lower the required minimum allocation given in the EXE header to the minimally required value so that decompression will still succeed. This defaults to using dosemu2, which must have a DOS installed that allows filesystem redirection. DEFAULT_MACHINE can be used to select qemu instead. The options BOOT_KERNEL, BOOT_COMMAND, and BOOT_PROTOCOL must be set up then to allow building a bootable diskette. (This is needed because qemu does not offer filesystem redirection for DOS.)
use_build_revision_id option is by default on. It requires that the sources are in hg (Mercurial) repos and that the hg command is available to run ‘hg id’. The resulting revision IDs are embedded into the executable and will be shown for the ?B (long) and ?BUILD (short) commands.
$NASM specifies the nasm executable to use, with path if needed.
./makec’ (or use whatever C compiler to build mktables) then ‘./mktables’ next. Note that mktables only needs to be used if either the source files (instr.*) changed or the mktables program itself has been altered. If the assembler and disassembler tables are not to change then mktables need not be used.
./mak.sh’ from the ldebug/source directory. You may pass environment variables to it, such as ‘INICOMP_METHOD=exodecr ./mak.sh’ to select Exomizer compression. You may also pass it parameters which will be passed to the main assembly command, such as ‘./mak.sh -D_DEBUG4’ to enable debugging messages.
The mak.sh script expects that the current working directory is equal to the directory that it resides in. So you'll always want to run it as ‘./mak.sh’ from that directory. The same is true of the make* scripts.
The make* scripts work as follows:
ldebug/tmp, ldebug/lst, and ldebug/bin will receive the files created by the mak script. The following filenames are for the default when running mak.sh on its own which is to create debug. (When ddebug, debugx, or ddebugx are created, the names change accordingly.) In the ldebug/bin subdirectory, debug.com will be a nonbootable executable (even if the _BOOTLDR option is enabled). This executable can safely be compressed using EXE packers such as the UPX. (In cfg.sh the option use_build_shim now controls whether debug.com is created. It defaults to disable this output file.) If the _BOOTLDR option is enabled, ldebug.com will be a compressed bootable executable (if any compression method is selected), whereas ldebugu.com will be an uncompressed bootable executable. These bootable executables must not be compressed using any other programs. Doing that would render the kernel mode entrypoints unusable. Incidentally, UPX rejects these files because their ‘last page size’ MZ EXE header field holds an invalid value.
The bootable executables can be used as MS-DOS 6 protocol IO.SYS, MS-DOS 7/8 IO.SYS, PC-DOS 6/7 IBMBIO.COM, FreeDOS KERNEL.SYS, RxDOS.3 RXDOS.COM, or as a Multiboot specification or Multiboot2 specification kernel. In any kernel load protocol case, the root FS that is being loaded from should be a valid FAT12, FAT16, or FAT32 file system on an unpartitioned (super)floppy diskette (unit number up to 127) or MBR-partitioned hard disk (unit number above 127). In addition, the bootable executables also are valid 86-DOS application programs that can be loaded in EXE mode. (Internally, all the .com files are MZ executables with a header, but they are named with a .COM file name extension for compatibility.)
It is valid to append additional data, such as a .ZIP archive, to any of the executables. However, if too large this may render loading with the FreeDOS load protocol impossible. All the other protocols work even in the presence of arbitrarily large appended data.
hg pull’ to update the repo
hg up default’ or the symbolic branch with ‘hg up symbolic’ or any other available commit you want to build
hg fetch’ or the sequence of ‘hg pull’ then ‘hg up’ to update the repos. (Usually the additional source repos do not have multiple branches.)
$NASM specifies the nasm executable to use, with path if needed.
./makinst.sh’ from the ldebug/source directory. You may pass environment variables to it. You may also pass it parameters which will be passed to the assembly commands.
The makinst.sh script expects that the current working directory is equal to the directory that it resides in. So you'll always want to run it as ‘./makinst.sh’ from that directory.
ldebug/tmp, ldebug/lst, and ldebug/bin will receive the files created by the makinst script. ldebug/bin/instsect.com will be the instsect application, which has boot sector loaders for FAT12, FAT16, and FAT32 embedded. The default protocol is lDOS and the default kernel name LDEBUG.COM. Read the instsect help page for instructions on how to use it. The help can be obtained by running instsect.com /? from DOS. The kernel name can be modified with the/F= switch to instsect. For instance, ‘instsect.com /f=lddebugu.com a:’ installs the loader onto drive A: with the name set up to load the uncompressed lDDebug.
Current lDOS boot32 uses the FSIBOOT4 protocol for an additional stage. This is interoperable with the upcoming RxDOS version 7.25's use of the FSIBOOT4 protocol, as well as with loaders that use a different sector for their additional stage (like Microsoft's), or those that do not use an additional stage (like FreeDOS's).
~’.
(To debug everything including the section from run to intrtn1_code, or the DPMI entry of lDebugX, a lower-level debugger must be used, such as dosemu's dosdebug or other debuggers that are integrated into emulators.)
-’ to a hash sign ‘#’. (lDDebugX prepends its tilde to that resulting in ‘~#’.)
ldebugu.com and ldebug.com. In bootloaded mode, I/O is never done using DOS, as if InDOS mode was always on. The DOS's current PSP is not switched during debugger operation. The MCB chain can only be displayed using the DM command by specifying the start segment explicitly. The BOOT commands are supported, refer to section 7.10.
The debugger can be loaded as a variety of kernel formats.
The Multiboot1 and Multiboot2 entrypoints will expect that a kernel command line is provided. The RxDOS.3 and lDOS load protocols allow specifying a kernel command line, but it is optional.
If a kernel command line is detected then its contents are entered into the command line buffer. Unescaped semicolons are translated into Carriage Returns. Semicolons and backslashes may be escaped with backslashes.
If no kernel command line is given, the debugger assumes a default. It is equivalent to checking for a file and label using the IF command (section 5.18), then if found to execute that script file. The IF condition is like if exists y ldp/LDEBUG.SLD :bootstartup then and the subsequent script command is y ldp/LDEBUG.SLD :bootstartup (section 5.39). The filename is however LDDEBUG.SLD for DDebug builds.
Executing the Q command (section 5.26) makes the debugger uninstall itself then continue running whatever code the debuggee is in. Executing the BOOT QUIT command (section 7.10) makes the debugger attempt to shut down the machine. First it will try to call a dosemu-specific callback. Next it will attempt shutting down with APM. (This works in qemu.) Finally it will give up if no attempt worked.
The debugger is internally an MZ .EXE style application. It may need MS-DOS version 3 level features. A few switches are supported:
/?
/c
/z
After the switches a filename may follow. After the filename, command line contents for the process to be debugged may follow. These are both passed to the N command. Then, an L command for loading an application is run.
Executing the Q command (section 5.26) makes the debugger try to terminate the debuggee application and to then terminate itself. The debugger returns to whatever application called it.
If the TSR command (section 5.34) is used, the debugger patches the parent of the currently running application to be the debugger's parent. A subsequent Q command will then behave much like it does in boot loaded mode: The debugger uninstalls itself and continues execution in the current debuggee context.
Plain numbers are evaluated as expressions. Expressions consist of any number of the following:
Plain number parsing for an expression continues for as long as a valid expression is continued. For example, in the command ‘D 100 + 20 L 10’ the starting address (its offset to be specific) is calculated as ‘100 + 20’. Then the expression evaluator encounters the ‘L’, which is not a valid binary operator. Plain number expression parameters are used by a lot of commands. Sometimes, the plain number parameter type is called ‘count’ or ‘value’.
An address parameter is calculated with a default segment. First, a plain number is parsed. If it is followed by a colon, the first number is taken as segment, and then another number is parsed for the offset. Otherwise, the first number is used as the offset. Offsets may be 16 bits or 32 bits wide, though 32-bit offsets are only valid for DebugX and only in 32-bit segments. Address parameters are used by a lot of commands.
A range parameter may have a default length, or it may be disallowed to omit a length. Parsing a range starts with parsing an address. Then, if the end of the line is not yet reached, an end for the range may be specified. The end may be a plain number, which is taken as the offset of the last byte to include in the range. The address of the last byte to include must be equal or above the address of the first byte that is included in the range.
The end may instead be specified with an ‘L’ or ‘LENGTH’ keyword. In that case, the keyword is followed by a plain number and an optional item size keyword. A length of zero is not valid. The item size keyword may be ‘BYTES’, ‘WORDS’, or ‘DWORDS’. For the latter two, the plain number will be multiplied by 2 or 4. The ‘BYTES’ keyword is only provided for symmetry; currently all commands taking ranges default to byte size for the ‘LENGTH’ number.
For example, the command ‘DD 100 LENGTH 4 DWORDS’ will dump memory from address 0100h (in the current data segment) in dword units, for a length of 4*4 = 16 bytes. The item size keywords were introduced primarily for the ‘DW’ and ‘DD’ commands (refer to section 5.8), but they can be used for any command that accepts a range.
Range parameters are used by a lot of commands.
A list is made up of a sequence of items. Each item is either a plain number or a quoted string. List parsing continues until the end of the line. Each plain number represents a single byte. Quoted strings represent as many bytes as there are quoted. List parameters are used by the E, F, and S commands. Refer to section 5.12, section 5.13, and section 5.31.
A list or range can be specified for this parameter. The range is identified by a leading ‘RANGE’ keyword. Otherwise, a list is parsed. A list or range parameter is as yet used by the S command and the F command, refer to section 5.31 and section 5.13.
A keyword is checked insensitive to capitalisation. Keywords depend on each command. Only the keywords used to specify a range's length are shared by all commands that parse ranges.
An index is a plain number that specifies a breakpoint index. It allows operating on one specific breakpoint. The index parameter type is used by the B commands, refer to section 5.5.
A segment is a plain number for parsing purposes. The segment parameter type is used by the DM command and some BOOT commands, refer to section 5.10 and section 7.10.
Each breakpoint is a single address, which defaults to the code segment. The address may instead be specified starting with an AT sign ‘@’, followed by a blank or an opening parenthesis. In that case, the following plain number specifies the non-segmented linear address to use. The breakpoint parameter type is used by the B and G commands, refer to section 5.5 and section 5.14.
A label is a (not quoted) string keyword. It may start with an optional colon. A label can be used by the GOTO and Y commands, refer to section 5.15 and section 5.39.
A port is a plain number for parsing purposes. The port parameter type is used by the I and O commands, refer to section 5.17 and section 5.24.
A drive may be either an alphabetic letter followed by a colon, or a plain number. The number zero corresponds to drive A: then. The drive parameter type is used by the L and W sector commands, refer to section 5.20 and section 5.37. The N and Y commands (section 5.23 and section 5.39) also accept drive parameters, but only as part of their filenames. These must be in the drive letter followed by colon format.
A sector is a plain number, which can be equal to any 32-bit value. The sector parameter type is used by the L and W sector commands, refer to section 5.20 and section 5.37. Some BOOT commands also use sector numbers, refer to section 7.10.
A condition is a plain number. It is evaluated either to nonzero (true) or zero (false). The condition parameter type is used by the IF command, as well as the P, TP, and T commands when specified with a ‘WHILE’ keyword. The BW and BP (with a ‘WHEN’ keyword) commands also use conditions. Refer to section 5.18, section 5.25, section 5.32, section 5.5.3, section 5.5.1. The length of a condition for B commands is limited by how much space is left in the permanent breakpoint conditions buffer. This buffer currently defaults to 1024 bytes. It is shared for all conditions of all permanent breakpoints.
A register specifies an internal variable of the debugger. Most prominently these include the debuggee's registers as stored by the debugger in its data segment. A register or variable may be an operand in a plain number's expression. However, several forms of the R command also use register parameters. These allow reading and writing the register values. Refer to section 5.27.
Command is a special parameter type that is used only by the RE.APPEND and RE.REPLACE commands (section 5.27.2). It is read verbatim and entered into the RE command buffer. Semicolons within a command parameter are not parsed as end of line comment markers. Instead, they are converted to CR (13) codes in the RE buffer. This delimits the parts of the parameter into several commands. A semicolon may be prefixed by a backslash to escape it and thus enter a literal semicolon into the RE buffer.
ID is a special parameter type that is used only by the BP and BI commands (section 5.5.1 and section 5.5.2). Leading and trailing whitespace is ignored. An ID can be empty, or contain up to 63 bytes of data. The length of an ID is also limited by how much space is left in the permanent breakpoint ID buffer. This buffer currently defaults to 384 bytes. It is shared for all IDs of all permanent breakpoints.
Entering an empty command at an interactive prompt results in autorepeat. Interactive prompts for this purpose include:
int 21h)
int 16h/int 10h)
Input that does not count as an interactive prompt includes:
int 21h)
int 21h)
int 13h)
Autorepeat is not supported by all commands. The following commands support autorepeat:
CS:(E)IP at the start of the command's execution, it is skipped once.
Online help ?
The question mark command (?) lists the main online help screen.
There are additional help topics that can be listed by using the question mark command with an additional letter or keyword. These keywords are as follows:
Registers ?R
Flags ?F
Conditionals ?C
Expressions ?E
Variables ?V
R Extended ?RE
Run keywords ?RUN
Options ?O
Boot loading ?BOOT
lDebug build ?BUILD
lDebug build ?B
lDebug sources ?SOURCE
lDebug license ?L
The full help pages are listed in section 7.
A leading colon indicates a destination label for GOTO, see section 5.15.
assemble A [address]
Starts assembly at the indicated address (which defaults to CS segment), or if no address is specified, at the "a_addr" (AAS:AAO variables).
Assembly mode has its own prompt. Entering a single dot (.) or an empty line terminates assembly mode. Comments can be given with a prefixed semicolon. In assembly mode, whereever an immediate number occurs an expression can be given surrounded by parentheses ( and ). In such expressions, register names like AX are evaluated to the values held by the registers at assembly time. To refer to a register as an assembly operand, it must occur outside parentheses.
There are a fixed number of permanent breakpoints provided by the debugger. The default is to provide 16 permanent breakpoints. They are specified by indices ranging from 00 to 0F. A breakpoint can be unused, used while enabled, or used while disabled. A breakpoint that is in use has a specific linear address. It is allowed, though not advised, for several breakpoints to be set to the same address.
When running the debuggee with the commands G, T, TP, or P, hitting a permanent breakpoint stops execution, and indicates in a message "Hit permanent breakpoint XX" where XX is replaced by the hexadecimal byte index of the breakpoint. If the breakpoint counter is not equal to 8000h when the breakpoint is hit, then the "Hit" message is followed by a "counter=YYYY" indicator. If the breakpoint ID is not empty, then the ID is shown with an "ID: " prefix. The ID is shown either on the same line as the "Hit" message, or on the next line if the ID exceeds 28 bytes. After that message a register dump occurs, same as for default breaking for the Run commands.
The exceptions are as follows:
Each breakpoint has a breakpoint counter, which defaults to 8000h if not set explicitly by the BP or BN commands. The breakpoint counter behaves as follows:
The point being passed means that during running the debuggee with a Run command, execution is not stopped, but a message indicating "Passed permanent breakpoint XX, counter=YYYY" is displayed. As for the "Hit" message the ID, if any, is also shown. After that message, a register dump occurs. Then execution is continued in accordance with the command that is running debuggee code.
Each breakpoint can have a breakpoint condition. If the condition expression evaluates to false when the point breaks, then the point is not considered hit or passed. The breakpoint counter is not stepped then either.
set breakpoint BP index|AT|NEW address
[[NUMBER=]number] [WHEN=cond] [ID=id]
BP initialises the breakpoint with the given index. It must be a yet unused breakpoint. If the index is specified as the keyword NEW, the lowest unused breakpoint (if any) is selected. If there is the keyword AT instead of an index or a keyword NEW, then an existing breakpoint at the same linear address is reset, or a new one is added (same as if given the NEW keyword).
The address can be given in a segmented format, which defaults to CS, and which in DebugX is subject to either PM or 86M segmentation semantics depending on which mode the debugger is in. The address can also be given with an @ specifier (followed by an opening parenthesis or whitespace) in which case it is specified as the 32-bit linear address. Debug without DPMI support limits breakpoints to 24-bit addresses, of which 21 bits are usable.
The optional number, which defaults to 8000h, sets the breakpoint counter to that number.
The optional WHEN keyword introduces a breakpoint condition. If the breakpoint is reached then the condition, if specified, is checked before stepping the counters. If the condition is false at that point the point is not considered hit or passed and its counter is not stepped.
There is an optional OFFSET keyword (not shown in the example) which allows overriding the breakpoint's preferred offset. Refer to section 5.5.4 for details.
The optional ID keyword allows setting the breakpoint ID. The ID is displayed by BL and when a breakpoint is hit or passed. The default ID is an empty ID. Note that the ID extends for the remainder of the line. There cannot be a breakpoint counter number nor WHEN condition nor OFFSET after the ID keyword.
set ID BI index|AT address [ID=]id
BI sets the breakpoint ID of the specified breakpoint. The ID is displayed by BL and when a breakpoint is hit or passed. The ID may be specified as empty.
set condition BW index|AT address [WHEN=]cond
The BW command sets the breakpoint condition. If the WHEN keyword and the condition are absent then the condition is reset. That means the point is no longer conditional.
set offset BO index|AT address [OFFSET=]number
The BO command sets the breakpoint preferred offset. The preferred offset is used only by the BL command. It is used to determine the segmented address to display. The offset is a word variable for Debug and a dword variable for DebugX. If the OFFSET keyword and the number are absent then the offset is disabled, as if the breakpoint was specified with a linear address. (Internally this is done by setting the offset to all 1 bits. The offset can be explicitly set to FFFFh (Debug) or FFFF_FFFFh (DebugX) for the same effect.)
set number BN index|AT address|ALL number
BN sets the breakpoint counter of the specified breakpoint with the given index, or all used breakpoints when given the keyword ALL, or the first breakpoint with a matching linear address when given the AT keyword. The number defaults to 8000h.
clear BC index|AT address|ALL
BC clears the specified breakpoint with the given index, or all breakpoints when given the keyword ALL, or the first breakpoint with a matching linear address when given the AT keyword. This returns the specified breakpoint (or all of them) to the unused state. Any associated ID or condition is deleted by BC too.
disable BD index|AT address|ALL
Given an index or the keyword ALL or the keyword AT (like BC), BD disables breakpoints that are in use. A disabled breakpoint's address is retained and BP will not allow initialising it anew (except with AT), but it is otherwise skipped in breakpoint handling.
enable BE index|AT address|ALL
Like BD, but enables breakpoints.
toggle BT index|AT address|ALL
Like BE and BD, but toggles breakpoints: A disabled breakpoint is enabled, while an enabled breakpoint is disabled.
list BL [index|AT address|ALL]
BL lists a specific breakpoint given by its index, or all used breakpoints if given the keyword ALL or given neither an index nor the keyword. When given the AT keyword, all breakpoints with a matching linear address are listed. (This differs from all other B commands, which only select the first matching breakpoint when the AT keyword is given.)
When listing all breakpoints only used breakpoints are displayed.
The output format for unused breakpoints is as follows:
The output format for used breakpoints is as follows:
Example output of BL:
-bp at 100 id = start
-bp at 103 counter = 4000
-bp at 105 when al == 7
-bl
BP 00 + Lin=01_BB70 1BA7:0100 (CC) Counter=8000, ID: start
BP 01 + Lin=01_BB73 1BA7:0103 (CC) Counter=4000
BP 02 + Lin=01_BB75 1BA7:0105 (CC) Counter=8000
WHEN al == 7
-
break upwards BU
This command, which is only supported by Debuggable lDebug builds (DDebug), causes the debugger to execute an int3 instruction in its own code segment. This breaks to the next debugger that was installed prior to DDebug. Prior to the breakpoint, the message "Breaking to next instance." is displayed.
In non-debuggable lDebug builds, the following error message is displayed instead:
-bu
Already in topmost instance. (This is no debugging build of lDebug.)
-
compare C range address
Given a range, the address of which defaults to DS, and another address that also defaults to DS, this command compares strings of bytes, and lists the bytes that differ.
dump D [range]
dump bytes DB [range]
dump words DW [range]
dump dwords DD [range]
Given a range, the address of which defaults to DS, this command dumps memory in hexadecimal and as ASCII characters. If the DCO option 4 is set, characters with the high bit set (80h to FFh) are displayed as-is in the character dump. Otherwise, they will be treated like control characters, which means replaced by dots.
If no range is specified, the D command continues dumping at "d_addr" (ADS:ADO), which is updated by each D command to point after the last shown byte.
The default is for D to dump bytes. After a DW or DD command, the autorepeat and plain D (without a range) default to the last-used size. If the default range should be used but the size should be reset to bytes, the DB command can be used. The D command with a range always acts the same as DB.
dump interrupts DI interrupt [count]
The DI command dumps interrupt vectors from the IVT (86M) or IDT (PM). In PM, for the vectors 00h to 1Fh, the exception handlers are also dumped.
dump MCB chain DM [segment]
The DM command dumps an MCB chain. If not given a start MCB segment, and the debugger is running as an 86-DOS application, the start of DOS's MCB chain is used. If given a start MCB segment, this is used as the starting MCB. (Note: In current RxDOS builds, the start MCB is always at segment 60h.)
The DM command initially lists the debuggee's PSP. This is only valid when the debugger is running as an 86-DOS application.
The MCB chain dump is continued until an MCB is encountered that has neither an M nor a Z signature letter, or the MCB address wraps around the 1 MiB boundary. In particular, this means that a disabled UMB link MCB (usually pointing to the MCB at segment 9FFFh if there is no EBDA nor any pre-boot-loaded programs) will not end the dump.
Example output:
-dm
PSP: 1A73
02B4 4D 0008 0016 352 B SD
02CB 4D 02CC 00BC 2 KiB COMMAND
0388 4D 039D 0013 304 B SYSTEM
039C 4D 039D 0034 832 B SYSTEM
03D1 4D 04A3 0013 304 B LDEBUG
03E5 4D 03E6 00BC 2 KiB COMMAND
04A2 4D 04A3 15CF 87 KiB LDEBUG
1A72 5A 1A73 858C 534 KiB DEBUGGEE
9FFF 4D 0008 3100 196 KiB SC
D100 4D 0008 1EFF 123 KiB SC
F000 4D 02CC 0040 1024 B COMMAND
F041 4D 0000 0492 18 KiB
F4D4 4D 0000 0619 24 KiB
FAEE 4D 0000 0090 2 KiB
FB7F 5A 03E6 0080 2048 B COMMAND
-
The columns are as follows:
M’) for linking MCB and 5A (‘Z’) otherwise.
SC/SD/S system MCB owner. Higher values are generally process segments. A process segment is usually a memory block that is preceded by an MCB, which is owned by that block itself.
display strings DZ/D$/D[W]# [address]
The D string commands each dump a string at a specified address, which defaults to DS as the segment.
enter E address [list]
fill F range [RANGE range|list]
The F command fills memory with a byte pattern. The first parameter is the range to fill. The next parameter can be a list, in which case it provides the pattern with which to fill. If the RANGE keyword is provided then the pattern is read from memory as indicated by the range parameter that follows the keyword. The pattern is repeated so as to fill the destination. If the RANGE keyword is used, then the length of the pattern address range is optional. If the length is absent, it is assumed to equal that of the destination range.
go G [=address] [breakpts]
The G command runs the debuggee. It can be given a start address (the segment of which defaults to CS), prefixed by an equals sign, in which case CS:EIP is set to that start address upon running. Note that if there is an error parsing the command line, CS:EIP is not changed. Further, if a breakpoint fails to be written initially, CS:EIP also is not changed.
The G command allows specifying breakpoints, which are either segmented addresses (86M or PM addresses depending on DebugX's mode) or linear addresses prefixed by an "@ " or "@(", similar to how the BP command allows a breakpoint specification. G breakpoints are identified by their position in the command line, as the 1st, 2nd, 3rd, etc. By default, 16 G breakpoints are supported.
The G AGAIN command re-uses the breakpoints given to the last (successfully parsed) G command. It also allows an equals-sign-prefixed start address like the plain G command, in front of the AGAIN keyword. After the AGAIN keyword, additional breakpoints may be specified.
If the command repetition of G is used, it is handled as if "G AGAIN" was entered, that is it re-uses the same breakpoints as those given to the prior G command.
A G command that fails to parse will not modify the stored G breakpoint list. If an error occurs during writing breakpoints, the list will have been modified already however.
The G LIST command lists the breakpoints given to the last (successfully parsed) G command.
The "content" byte in G LIST is usually CCh (the int3 instruction opcode), but retains its original value if a failure occurs during breakpoint byte restoration.
Example output of G LIST:
-g 100 103 105
AX=3000 BX=0000 CX=0200 DX=0000 SP=FFFE BP=0000 SI=0000 DI=0000
DS=1BA7 ES=1BA7 SS=1BA7 CS=1BA7 IP=0103 NV UP EI PL ZR NA PE NC
1BA7:0103 CD21 int 21
-g list
1st G breakpoint, linear 0001_BB70 1BA7:0100, content CC
2nd G breakpoint, linear 0001_BB73 1BA7:0103, content CC (is at CS:IP)
3rd G breakpoint, linear 0001_BB75 1BA7:0105, content CC
-
The output is as follows:
CS:’ if the code segment's base matches the preferred offset. Otherwise, an R86M segment is shown with a dollar sign ‘$’ prefix if the preferred offset matches any R86M segment. Failing that the offset is shown with a prefix reading ‘????:’.
There is another G command: After any equals sign, AGAIN keyword, and/or specified breakpoints, the line can be ended with a REMEMBER keyword. This saves the specified G breakpoint list and then returns control to the user. (The equals address, if any, is discarded.) It allows preparing a G breakpoint list ahead of its use. Auto-repeat, if enabled, will run like G AGAIN and actually run the debuggee after a G REMEMBER command.
goto GOTO :label
The GOTO command can only be used when executing from a script file, the command line buffer, or the RE buffer. It lets execution continue at a different point in the file or buffer. Labels are identified by lines that start with a colon, followed by the alphanumeric label name, and optionally followed by a trailing colon. The destination label of the GOTO command may be specified with or without the leading colon.
There are several special cases:
hex add/sub H value1 [value2 [...]]
The H command performs calculation and displays the result. If a single expression is given then its value is displayed, in hexadecimal and then in decimal. If more than one expression is given then two results are displayed, in hexadecimal only. The first result is that which is calculated by adding all expressions. The second result is calculated by subtracting all subsequent expressions from the first expression's value.
If a value is above or equal to 8000_0000h then along each display of that value, the value interpreted as a negative two's complement number is listed in parentheses.
Examples:
-h 1
0001 decimal: 1
-h 1 1
0002 0000
-h 1 1 1
0003 FFFFFFFF (-0001)
-h 1 + 2 * 3
0007 decimal: 7
-h cs * 10
0001A730 decimal: 108336
-h -26
FFFFFFDA (-0026) decimal: 4294967258 (-38)
-
input I[W|D] port
The I commands input from an x86 port. The port can be any number between 0 and FFFFh. Plain I inputs a byte from the specified port. The IW and ID commands input a word or dword respectively.
if numeric IF [NOT] (cond) THEN cmd
if script file IF [NOT] EXISTS Y file [:label] THEN cmd
The IF command allows specifying a conditionally executed command. This is especially useful for creating conditional control flow branches with the GOTO command (see section 5.15).
For the first form, the condition is a numeric expression. If it evaluates to non-zero it is considered true. If the NOT keyword is absent then a true condition expression leads to executing the THEN command. With the NOT keyword present the logic is reversed. Note that if an error occurs in parsing, the THEN command is not executed, regardless of whether the NOT keyword is present.
The second form specifies a script file in the same format as accepted by the Y command (refer to section 5.39). A label may be specified behind the filename, as for the Y command. If the file is found, and contains the specified label if any, then the EXISTS clause is considered true. Depending on the presence of the NOT keyword the THEN command is executed next, or skipped. Note that if an error occurs in parsing, the THEN command is not executed, regardless of whether the NOT keyword is present.
Likewise, if an unanticipated error occurs during access then the THEN command is not executed. Anticipated errors include:
load program L [address]
load sectors L address drive sector count
move M range address
80x86/x87 mode M [0..6|C|NC|C2|?]
set name N [[drive:][path]progname.ext [parameters]]
output O[W|D] port value
The O commands output to an x86 port. The port can be any number between 0 and FFFFh. Plain O outputs a byte to the specified port. The OW and OD commands output a word or dword respectively. The value to write is specified by the second expression.
proceed P [=address] [count [WHILE cond] [SILENT [count]]]
The P command causes debuggee to run a proceed step. This is the same as tracing (T command) for most instructions, but behaves differently for "call", "loop", and repeated string instructions. For these, a proceed breakpoint is written behind the instruction (similarly to how the G command writes breakpoints), and the debuggee is run without the Trace Flag set.
Like for the G command, a start address can be given to P prefixed by an equals sign. Next, a count may be specified, which causes the command to execute as many P steps as the count indicates.
After a count, a WHILE keyword may be specified, which must be followed by a conditional expression. Execution will only continue if the WHILE expression evaluates to true.
After a count (when no WHILE is given) or after a WHILE condition, a SILENT keyword and optional count may be given. In this case, the debugger buffers the register dump and disassembly output of the executed steps, until control returns to the debugger command line. Then, the last dumps stored in the buffer are displayed. If a non-zero count is given, at most that many register dumps are displayed.
quit Q
register R [register [value]]
The R command without any register specified dumps the current registers, either displayed as 16-bit or 32-bit values (depending on the RX option), and disassembles the instruction at the current CS:(E)IP location.
R with a register, named debugger variable, or memory variable (of the form BYTE/WORD/3BYTE/DWORD [segment:offset]) displays the current value of the specified variable. It then displays a prompt, allowing the user to enter a new value for that variable. Entering a dot (.) or an empty line returns to the default debugger command line.
R with a variable, followed by a dot (.), only displays the current value of that variable.
R with a variable, followed by an optional equals sign, and followed by an expression, evaluates the expression and assigns its resulting value to the variable. The equals sign may instead be a binary operator with a trailing equals sign, which is handled as an assignment operator.
Examples:
-r ax .
AX 0000
-r ax
AX 0000 :1
-r ax
AX 0001 :.
-r ax += 4
-r ax
AX 0005 :
-r word [cs:0]
WORD [1867:0000] 20CD :
-r dif .
DIF 0100B00B
-
R extended RE
The RE command runs the RE buffer commands. Refer to section 7.7.
RE commands RE.LIST|APPEND|REPLACE [commands]
RE.LIST lists the RE buffer contents in a way that can be re-used as input to RE.REPLACE.
RE.APPEND appends the following commands to the RE buffer.
RE.REPLACE replaces the RE buffer with the following commands.
The RE buffer usage is described in the ?RE help page (section 7.7).
MMX register RM
FPU register RN
toggle 386 regs RX
search S range [REVERSE] [RANGE range|list]
The S command searches memory for a byte string. The first range specifies the search space. By default, searching will begin at the bottom of the search space and move upwards. If a REVERSE keyword is specified after the range then searching will begin at the top of the search space moving downwards. The search string is specified either with the RANGE keyword followed by another range, or as a list of byte values.
The read-only variable SRC (Search Result Count) will receive the 32-bit value that is the amount of matched occurrences. The variable SRS0 receives the first Search Result Segment. Likewise SRO0 receives the first Search Result Offset. SRO1 to SROF hold subsequent Search Result Offsets. SRO is an alias to SRO0. SRO variables are 32-bit in the _PM build lDebugX, 16-bit otherwise. Unused SRO variables are zeroed out by a successful search.
The display of search results is as follows:
There is an option to disable the data dump so as to only display the match addresses. If the bit 80_0000h is set in the DCO variable then the data dump is suppressed.
trace T [=address] [count [WHILE cond] [SILENT [count]]]
The T command is similar to the P command. However, T traces most instructions. Depending on the TM option, interrupt instructions are also traced (into the interrupt handler) or proceeded past.
trace (exc str) TP [=address] [count [WHILE cond] [SILENT [count]]]
The TP command is alike the T command, but proceeds past repeated string instructions like the P command would.
trace mode TM [0|1]
enter TSR mode TSR
unassemble U [range]
write program W [address]
write sectors W address drive sector count
expanded mem XA/XD/XM/XR/XS, X? for help
run script Y [partition/][scriptfile] [:label]
The Y command runs a script file. The script file is specified in two different ways, depending on whether the debugger is running as an 86-DOS application or as a boot-loaded kernel replacement.
Further, a label may be specified to cause execution to start at that label instead of at the start of the file. This is equivalent to placing a ‘GOTO :label’ command at the start of the script file. The colon to indicate a label is required.
If execution already is within a script file, then the Y command may be run with only a label (again with the colon required). In that case, the current script file is opened in a subsequent level (handle or boot-loaded script file context) and execution starts at that label.
All debuggee registers can be accessed numerically:
al, cl, dl, bl, ah, ch, dh, bh
ax, cx, dx, bx, sp, bp, si, di
eax, ecx, edx, ebx, esp, ebp, esi, edi
es, cs, ss, ds, fs, gs
fl, efl, ip, eip
Each 16-bit register can be used in a register pair, such as:
dxax
bxcx (used by L load program and W write program commands)
sidi
csip
0 while in Real or Virtual 8086 Mode, debugger process selector otherwise. (The process selector addresses DebugX's PSP and DATA ENTRY section.)
All of these are doublewords and default to 1. For the respective commands, these counts specify the number of steps to take if none is specified explicitly. This includes when a command is run by autorepeat, refer to section 5.1. If one of these is set to zero then it is an error to not specify a count explicitly for the corresponding command.
Doubleword. Default is 256. If this many commands are executed from the RE buffer, the execution is aborted and the command that called RE is continued.
Doubleword. This is reset to zero when RE buffer execution starts. Each time a command is executed from the RE buffer, this variable is incremented. If it reaches the value of RELIMIT, RE buffer execution is aborted.
Word. This holds the most recent command's return code. If the most recent command succeeded, then this is zero.
Word. This holds the most recent non-zero return code.
AAS: word, AAO: doubleword. Default address for the assembler. Updated to point after each assembled instruction.
Default address for memory dumping. Updated to point after each dumped memory content.
Default address for the disassembler.
Default address for memory entry.
Default address for DZ command, ASCIZ strings. Terminated by zero byte.
Default address for D$ command, CP/M strings. Terminated by dollar sign ‘$’.
Default address for D# command, Pascal strings. Prefixed by length count byte.
Default address for DW# command. Prefixed by length count word.
Default address for DX command. (Only included in DebugX.)
Byte. Default 1. Sets the number of rows of the terminal used by DOS or BIOS output. Setting this to zero disables paging to the DOS or BIOS output. Setting this to 1 uses the automatic selection. That means the BIOS Data Area byte at address 484h, plus one, is used. If using that byte and it is zero, paging is disabled.
Byte. Default 1. Sets the number of columns of the terminal used by BIOS input. Setting this to zero selects a default (80). Setting this to 1 uses the automatic selection. That means the BIOS Data Area word at address 44Ah is used. This is used by the line input handling if inputting from the BIOS terminal (int 16h, int 10h), or if inputting from a DOS terminal when DCO flag 800h is set.
Byte. Default 24. Sets the number of rows of the terminal connected via serial port. Setting this to zero disables paging to the serial port. Setting this to 1 uses the IOR variable handling.
Byte. Default 80. Sets the number of columns of the terminal connected via serial port. Setting this to zero selects a default (80). Setting this to 1 uses the IOC variable handling. This is used by the line input handling.
Byte. Default 15. This gives the number of seconds that the KEEP prompt upon serial connection waits. Setting this to zero waits at the prompt forever.
Byte. Default 16. This gives the size of the 16550A's built-in TX FIFO to use. Set to 15 if using dosemu before revision gc7f5a828 2019-01-22, see https://github.com/stsp/dosemu2/issues/748.
Byte. Default 0Bh, corresponding to COM2. Use 0Ch for COM1. This specifies the interrupt number to hook so as to be notified of serial events. The use of this variable occurs only when connecting to serial I/O. The value at that point in time is cached for as long as the serial connection is in use.
Word. Default 0000_1000b, corresponding to COM2. Use 0001_0000b for COM1. This specifies the IRQ mask of which IRQs to enable. The low 8 bits correspond to IRQ #0 to #7 and the high 8 bits correspond to IRQ #8 to #15. If any bit of the high 8 bits is set then generally the bit 0100b should be set too, to enable the chained PIC. This circumstance is not automatically detected. The use of this variable occurs only when connecting to serial I/O. The value at that point in time is cached for as long as the serial connection is in use.
Word. Default 02F8h, corresponding to COM2. Use 03F8h for COM1. This specifies the I/O port base to address the UART. The use of this variable occurs only when connecting to serial I/O. The value at that point in time is cached for as long as the serial connection is in use.
Word. Default 12, corresponding to 9600 baud. This specifies the DL value to set during initialisation. The use of this variable occurs only when connecting to serial I/O.
Byte. Default 0000_0011b, corresponding to 8n1. (8n1 = 8 data bits, no parity, 1 stop bit.) This specifies the settings to set up in LCR. The high bit (80h) generally must be clear. The use of this variable occurs only when connecting to serial I/O.
Byte. Default 0. This specifies what to write to the FCR. The low 3 bits (07h) generally must be clear. The use of this variable occurs only when connecting to serial I/O. The value at that point in time is cached for as long as the serial connection is in use.
These variables are not supported by default. The build option _DEBUG1 must be enabled to include them. The Test Counter variables work similarly to permanent breakpoint counters:
The default values for all counters and addresses is zero.
If a fault is injected into readmem, it returns the value given in TRV.
If a fault is injected into writemem, it returns failure (CY).
If a fault is injected into getlinear, it returns failure (CY).
If a fault is injected into getsegmented, it returns failure (CY).
These variables are not supported by default. The build option _DEBUG3 must be enabled to include them. These variables are used to test the read-only masking. Read-only masking makes it so that bits given in the mask are read-only. Bits that are clear in the mask are writable.
Doubleword. Default 0. Mask AA55_AA55h.
Doubleword. Default 0011_0022h. Mask 00FF_00FFh.
Y command variables can be used when the Y command (as application or bootloaded) has been used to open a script file. YSx (Y Script) variables are generic and refer to whatever Y file is opened. YBx (Y Bootloaded script) variables refer to opened Y files while bootloaded. YHx (Y Handle script) variables refer to opened Y files as application.
Word. Partially read-write, partially read-only.
Flag 4000h controls whether script file input is displayed or not. Prepending an AT sign (@) to a line that is read from a script file will hide the input of that line. Setting YSF flag 4000h will hide all input lines instead. The effect is similar to prepending @ to every line.
YSF variables are only available while executing script files.
Doubleword. Default zero. V0 to VF or V00 to VFF each specify one variable. It is valid to refer to any V variable using an index expression. Index expression means that the variable name (V) is immediately followed by an opening parenthesis, followed by a numeric expression which evaluates to a number below 100h.
Doubleword. Read only. Amount of matches found by last S command.
Word. Read only. SRS0 to SRSF each specify one variable. Search result segments of last S command's matches.
Word or doubleword (DebugX). Read only. SRO0 to SROF each specify one variable. Search result offsets of last S command's matches. It is valid to refer to any SRO variable using an index expression. Index expression means that the variable name (SRO) is immediately followed by an opening parenthesis, followed by a numeric expression which evaluates to a number below 10h.
lDebug (2020-03-25) help screen
assemble A [address]
set breakpoint BP index|AT|NEW address
[[NUMBER=]number] [WHEN=cond] [ID=id]
set ID BI index|AT address [ID=]id
set condition BW index|AT address [WHEN=]cond
set offset BO index|AT address [OFFSET=]number
set number BN index|AT address|ALL number
clear BC index|AT address|ALL
disable BD index|AT address|ALL
enable BE index|AT address|ALL
toggle BT index|AT address|ALL
list BL [index|AT address|ALL]
compare C range address
dump D [range]
dump bytes DB [range]
dump words DW [range]
dump dwords DD [range]
dump interrupts DI interrupt [count]
dump MCB chain DM [segment]
display strings DZ/D$/D[W]# [address]
enter E address [list]
fill F range [RANGE range|list]
go G [=address] [breakpts]
goto GOTO :label
hex add/sub H value1 [value2 [...]]
input I[W|D] port
if numeric IF [NOT] (cond) THEN cmd
if script file IF [NOT] EXISTS Y file [:label] THEN cmd
load program L [address]
load sectors L address drive sector count
move M range address
80x86/x87 mode M [0..6|C|NC|C2|?]
set name N [[drive:][path]progname.ext [parameters]]
output O[W|D] port value
proceed P [=address] [count [WHILE cond] [SILENT [count]]]
quit Q
register R [register [value]]
R extended RE
RE commands RE.LIST|APPEND|REPLACE [commands]
MMX register RM
FPU register RN
toggle 386 regs RX
search S range [REVERSE] [RANGE range|list]
trace T [=address] [count [WHILE cond] [SILENT [count]]]
trace (exc str) TP [=address] [count [WHILE cond] [SILENT [count]]]
trace mode TM [0|1]
enter TSR mode TSR
unassemble U [range]
write program W [address]
write sectors W address drive sector count
expanded mem XA/XD/XM/XR/XS, X? for help
run script Y [partition/][scriptfile] [:label]
Additional help topics:
Registers ?R
Flags ?F
Conditionals ?C
Expressions ?E
Variables ?V
R Extended ?RE
Run keywords ?RUN
Options ?O
Boot loading ?BOOT
lDebug build ?BUILD
lDebug build ?B
lDebug sources ?SOURCE
lDebug license ?L
Available 16-bit registers: Available 32-bit registers: (386+)
AX Accumulator EAX
BX Base register EBX
CX Counter ECX
DX Data register EDX
SP Stack pointer ESP
BP Base pointer EBP
SI Source index ESI
DI Destination index EDI
DS Data segment
ES Extra segment
SS Stack segment
CS Code segment
FS Extra segment 2 (386+)
GS Extra segment 3 (386+)
IP Instruction pointer EIP
FL Flags EFL
Available 64-bit Matrix Math Extension (MMX) registers: (if supported)
MMx MM(x) MMX register x, where x is 0 to 7
Enter ?F to display the recognized flags.
Recognized flags:
Value Name Set Clear
0800 OF Overflow Flag OV Overflow NV No overflow
0400 DF Direction Flag DN Down UP Up
0200 IF Interrupt Flag EI Enable interrupts DI Disable interrupts
0080 SF Sign Flag NG Negative PL Plus
0040 ZF Zero Flag ZR Zero NZ Not zero
0010 AF Auxiliary Flag AC Auxiliary carry NA No auxiliary carry
0004 PF Parity Flag PE Parity even PO Parity odd
0001 CF Carry Flag CY Carry NC No carry
The short names of the flag states are displayed when dumping registers and can be entered to modify the symbolic F register with R. The short names of the flags can be modified by R.
In the register dump displayed by the R, T, P and G commands, conditional jumps are displayed with a notice that shows whether the instruction will cause a jump depending on its condition and the current register and flag contents. This notice shows either "jumping" or "not jumping" as appropriate.
The conditional jumps use these conditions: (second column negates)
jo jno OF
jc jb jnae jnc jnb jae CF
jz je jnz jne ZF
jbe jna jnbe ja ZF||CF
js jns SF
jp jpe jnp jpo PF
jl jnge jnl jge OF^^SF
jle jng jnle jg OF^^SF || ZF
j(e)cxz (e)cx==0
loop (e)cx!=1
loopz loope (e)cx!=1 && ZF
loopnz loopne (e)cx!=1 && !ZF
Enter ?F to display a description of the flag names.
Recognized operators in expressions:
| bitwise OR || boolean OR
^ bitwise XOR ^^ boolean XOR
& bitwise AND && boolean AND
>> bit-shift right > test if above
>>> signed bit-shift right < test if below
<< bit-shift left >= test if above-or-equal
>< bit-mirror <= test if below-or-equal
+ addition == test if equal
- subtraction != test if not equal
* multiplication => same as >=
/ division =< same as <=
% modulo (A-(A/B*B)) <> same as !=
** power
Implicit operater precedence is handled in the listed order, with increasing precedence: (Brackets specify explicit precedence of an expression.)
boolean operators OR, XOR, AND (each has a different precedence)
comparison operators
bitwise operators OR, XOR, AND (each has a different precedence)
shift and bit-mirror operators
addition and subtraction operators
multiplication, division and modulo operators
power operator
Recognized unary operators: (modifying the next number)
+ positive (does nothing)
- negative
~ bitwise NOT
! boolean NOT
? absolute value
!! convert to boolean
Note that the power operator does not affect unary operator handling. For instance, "- 2 ** 2" is parsed as "(-2) ** 2" and evaluates to 4.
Although a negative unary and signed bit-shift right operator are provided the expression evaluator is intrinsically unsigned. Particularly the division, multiplication, modulo and all comparison operators operate unsigned. Due to this, the expression "-1 < 0" evaluates to zero.
Recognized terms in an expression:
32-bit immediates
8-bit registers
16-bit registers including segment registers (except FS, GS)
32-bit compound registers made of two 16-bit registers (eg DXAX)
32-bit registers and FS, GS only if running on a 386+
32-bit variables V00..VFF
32-bit special variables DCO, DCS, DAO, DAS, DIF, DPI, PPI
16-bit special variables DPR, DPP, PSP, PPR
(fuller variable reference in the manual)
byte/word/3byte/dword memory content (eg byte [seg:ofs], where both the
optional segment as well as the offset are expressions too)
The expression evaluator case-insensitively checks for names of variables and registers as well as size specifiers.
Enter ?R to display the recognized register names. Enter ?V to display the recognized variables.
Available lDebug variables:
The following variables cannot be written:
Enter ?O to display the options and internal flags.
The RUN commands (T, TP, P, G) and the RE command use the RE command buffer to run commands. Most commands are allowed to be run from the RE buffer. Disallowed commands include program-loading L, A, E that switches the line input mode, TSR, Q, Y, RE, and further RUN commands. When the RE buffer is used as input during T, TP, or P with either of the WHILE or SILENT keywords, commands that use the auxbuff are also disallowed and will emit an error noting the conflict.
RE.LIST shows the current RE buffer contents in a format usable by the other RE commands. RE.APPEND appends the following commands to the buffer, if they fit. RE.REPLACE appends to the start of the buffer. When specifying commands, an unescaped semicolon is parsed as a linebreak to break apart individual commands. Backslashes can be used to escape semicolons and backslashes themselves.
Prefixing a line with an @ (AT sign) causes the command not to be shown to the standard output of the debugger when run. Otherwise, the command will be shown with a percent sign % or ~% prompt.
The default RE buffer content is @R. This content is also detected and handled specifically; if found as the only command the handler directly calls the register dump implementation without setting up and tearing down the special execution environment used to run arbitrary commands from the RE buffer.
T (trace), TP (trace except proceed past string operations), and P (proceed) can be followed by a number of repetitions and then the keyword WHILE, which must be followed by a conditional expression.
The selected run command is repeated as many times as specified by the number, or until the WHILE condition evaluates no longer to true.
After the number of repetitions or (if present) after the WHILE condition the keyword SILENT may follow. If that is the case, all register dumps done during the run are buffered by the debugger and the run remains silent. After the run, the last dumps are replayed from the buffer and displayed. At most as many dumps as fit into the buffer are displayed. (The buffer is currently up to 8 KiB sized.)
If a number follows behind the SILENT keyword, only at most that many dumps are displayed from the buffer. The dumps that are displayed are always those last written into the buffer, thus last occurred.
Available options: (read/write DCO, read DCS)
More options: (read/write DCO2, read DCS2)
More options: (read/write DCO3, read DCS3)
Internal flags: (read DIF)
Available assembler/disassembler options: (read/write DAO, read DAS)
Boot loading commands:
Available protocols: (default filenames, load segment, then entrypoint)
Available options:
Boolean options: [opt=bool]
lDebug (YYYY-MM-DD)
Source Control Revision ID: hg xxxxxxxxxxxx
Uses yyyyyyyy: Revision ID hg zzzzzzzzzzzz
[etc]
lDebug (YYYY-MM-DD)
Source Control Revision ID: hg xxxxxxxxxxxx
Uses yyyyyyyy: Revision ID hg zzzzzzzzzzzz
[etc]
DI command
DM command
D string commands
S match dumps line of following data
RN command
Access SDA current PSP field
Load NTVDM VDD for sector access
X commands for EMS access
RM command and reading MMX registers as variables
Expression evaluator
Indirection in expressions
Variables with user-defined purpose
Debugger option and status variables
PSP variables
Conditional jump notice in register dump
TSR mode (Process detachment)
Boot loader
Permanent breakpoints
Intercepted interrupts: 00, 01, 03, 06, 18, 19
Extended built-in help pages
Expanded memory (EMS) commands:
Allocate XA count
Deallocate XD handle
Map memory XM logical-page physical-page handle
Reallocate XR handle count
Show status XS
The original lDebug releases can be obtained from the repo located at https://hg.ulukai.org/ecm/ldebug (E. C. Masloch's repo)
The most recent manual is hosted at https://ulukai.org/ecm/doc/ in the files ldebug.htm, ldebug.txt, and ldebug.pdf
lDebug - libre 86-DOS debugger
Usage of the works is permitted provided that this instrument is retained with the works, so that any entity that uses the works is notified of this instrument.
DISCLAIMER: THE WORKS ARE WITHOUT WARRANTY.
All contributions by Paul Vojta or C. Masloch to the debugger are available under a choice of three different licenses. These are the Fair License, the Simplified 2-Clause BSD License, or the MIT License.
This is the license and copyright information that applies to lDebug; but note that there have been substantial contributions to the code base that are not copyrighted (public domain).
The program executables can be compressed with a choice of different compressors. The files then contain a decompression stub. Some of these stubs have their own usage conditions. The following stub usage conditions apply, if one of these stubs is used.
BriefLZ - small fast Lempel-Ziv
8086 Assembly lDOS iniload payload BriefLZ depacker
Based on: BriefLZ C safe depacker
Copyright (c) 2002-2016 Joergen Ibsen
This software is provided 'as-is', without any express or implied warranty. In no event will the authors be held liable for any damages arising from the use of this software.
Permission is granted to anyone to use this software for any purpose, including commercial applications, and to alter it and redistribute it freely, subject to the following restrictions:
Usage of the works is permitted provided that this instrument is retained with the works, so that any entity that uses the works is notified of this instrument.
DISCLAIMER: THE WORKS ARE WITHOUT WARRANTY.
Usage of the works is permitted provided that this instrument is retained with the works, so that any entity that uses the works is notified of this instrument.
DISCLAIMER: THE WORKS ARE WITHOUT WARRANTY.
Copyright (c) 2005-2017 Magnus Lind.
This software is provided 'as-is', without any express or implied warranty. In no event will the authors be held liable for any damages arising from the use of this software.
Permission is granted to anyone to use this software for any purpose, including commercial applications, and to alter it and redistribute it freely, subject to the following restrictions:
MIT License
Copyright (c) 2020 David Barina
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
Usage of the works is permitted provided that this instrument is retained with the works, so that any entity that uses the works is notified of this instrument.
DISCLAIMER: THE WORKS ARE WITHOUT WARRANTY.
This program is free software. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
Usage of the works is permitted provided that this instrument is retained with the works, so that any entity that uses the works is notified of this instrument.
DISCLAIMER: THE WORKS ARE WITHOUT WARRANTY.
hg 5f34bad78830, from commit on at 2021-02-15 12:26:03 +0100
If this is in ecm's repository, you can find it at https://hg.ulukai.org/ecm/ldebug/rev/5f34bad78830