VT-100命令工作奇怪

时间:2021-04-20 00:19:24

I'm writing simple clock program with NASM. I'm using Ubuntu 14.10 Vagrant box under OSX via iTerm. Terminal is xterm, so should be VT-100 compatible.

我正在用NASM编写简单的时钟程序。我通过iTerm在OSX下使用Ubuntu 14.10 Vagrant box。终端是xterm,因此应该兼容VT-100。

I need to remove one line. For instance, I expect the following behaviour:

我需要删除一行。例如,我期望以下行为:

Hello, this is clock program
13:01:25 UTC+4

Next moment:

Hello, this is clock program
13:01:26 UTC+4

I wrote the following functions. For print:

我写了以下函数。用于打印:

func_print:
  mov eax, sys_write
  mov ebx, stdout
  int 0x80
  ret

For clear:

clr          db 0x1b,  "[K"
clr_len      equ $-clr
...
func_clear:
  mov ecx, clr
  mov edx, clr_len
  call func_print

For saving and restoring positions I use VT-100 and its commands: [7 and [8 respectively:

为了保存和恢复位置,我使用VT-100及其命令:[7和[8]:

csave db     0x1b, "[7"
csave_len    equ $-csave

crestore     db 0x1b, "[8"
crestore_len equ $-crestore

My code:

global  _start
  _start:
    mov ecx, welcome
    mov edx, welcome_len
    call func_print

    call func_print
    call func_save_cursor_pos

    mov dword [tv_sec], 2
    mov dword [tv_usec], 0

    call func_sleep
    call func_clear

    call func_restore_cursor_pos
    mov ecx, welcome
    mov edx, welcome_len
    call func_print

    jmp func_exit

However, the result is:

但结果是:

vagrant@vagrant-ubuntu-trusty-64:~$ ./run.sh
Hello, this is the clock program
Hello, this is the clock program
Hello, this is the clock program
vagrant@vagrant-ubuntu-trusty-64:~$

If I change clr by adding [1A or [1B it seems to be removing lines to much higher than needed or lower:

如果我通过添加[1A或[1B]来改变clr,它似乎是将线条移除到远高于所需要的值或更低:

vagrant@vagrant-ubuntu-trusty-64:~$ ./run.sh
Hello, this is the clock program
Hello, this is the clock program








Hello, this is the clock program
vagrant@vagrant-ubuntu-trusty-64:~$

How can I fix it? What is the correct code?

我该如何解决?什么是正确的代码?

1 个解决方案

#1


I suspect that your issue is related to the newline implied to be in welcome db "Hello, this is the clock program", 10. I can't tell for sure because you didn't post that part of your code.

我怀疑你的问题与欢迎db中暗示的新行有关“你好,这是时钟程序”,10。我无法确定,因为你没有发布你的代码部分。

I think that this causes an issue because the newline causes the terminal to scroll - when I removed the newline from my version, it worked properly. If you only need to update one line, it works to just not have a newline.

我认为这会导致一个问题,因为换行导致终端滚动 - 当我从我的版本中删除换行符时,它正常工作。如果您只需要更新一行,它就可以没有换行符。

I suspect that the save and restore operations work on literal physical locations on the screen - not logical locations that get scrolled by newlines.

我怀疑保存和恢复操作在屏幕上的文字物理位置上工作 - 而不是由换行符滚动的逻辑位置。

In general, however, I'd recommend using the cursor manipulation escape codes instead:

但是,一般情况下,我建议使用光标操作转义码:

  • When you're ready to redraw your output, write db 0x1b, "[nA" to move up n lines. (You'll need to put the number in there.)
  • 当您准备重绘输出时,请写入db 0x1b,“[nA”向上移动n行。 (你需要把号码放在那里。)

  • Immediately after that (or after any subsequent newline), write db 0x1b, "[K" to clear that line. (You already knew this, but I'm including it for completeness.)
  • 紧接着(或任何后续换行符之后),写入db 0x1b,“[K”清除该行。 (你已经知道了这一点,但我将其包括在内是为了完整性。)

I wrote an example program to implement this, based partially on yours. It displays:

我写了一个示例程序来实现它,部分基于你的。它显示:

Hello, this is the clock program.
Line two.

Then, a short while later

不久之后

=== TEST ===
More.

And then

=== TEST 2 ===
Again.

This technique should be generalizable to any reasonable number of lines.

该技术应该可以推广到任何合理数量的线。

BITS 32
section .text

welcome      db "Hello, this is the clock program", 10, "Line two.", 10
welcome_len  equ $-welcome

test_str     db 0x1b, "[2A", 0x1b, "[K=== TEST ===", 10, 0x1b, "[KMore.", 10
test_len     equ $-test_str

test2_str     db 0x1b, "[2A", 0x1b, "[K=== TEST 2 ===", 10, 0x1b, "[KAgain.", 10
test2_len     equ $-test2_str

func_print:
    mov eax, 4
    mov ebx, 1
    int 0x80
    ret

pause: ; Note: DON'T EVER USE THIS IN A REAL PROGRAM. This is not how you sleep properly.
    mov eax, 0
loop:
    inc eax
    cmp eax, 1000000000
    jl loop
    ret

global  _start
_start:
    mov ecx, welcome
    mov edx, welcome_len
    call func_print

    call pause

    mov ecx, test_str
    mov edx, test_len
    call func_print

    call pause

    mov ecx, test2_str
    mov edx, test2_len
    call func_print

    mov eax, 1
    mov ebx, 0
    int 0x80

#1


I suspect that your issue is related to the newline implied to be in welcome db "Hello, this is the clock program", 10. I can't tell for sure because you didn't post that part of your code.

我怀疑你的问题与欢迎db中暗示的新行有关“你好,这是时钟程序”,10。我无法确定,因为你没有发布你的代码部分。

I think that this causes an issue because the newline causes the terminal to scroll - when I removed the newline from my version, it worked properly. If you only need to update one line, it works to just not have a newline.

我认为这会导致一个问题,因为换行导致终端滚动 - 当我从我的版本中删除换行符时,它正常工作。如果您只需要更新一行,它就可以没有换行符。

I suspect that the save and restore operations work on literal physical locations on the screen - not logical locations that get scrolled by newlines.

我怀疑保存和恢复操作在屏幕上的文字物理位置上工作 - 而不是由换行符滚动的逻辑位置。

In general, however, I'd recommend using the cursor manipulation escape codes instead:

但是,一般情况下,我建议使用光标操作转义码:

  • When you're ready to redraw your output, write db 0x1b, "[nA" to move up n lines. (You'll need to put the number in there.)
  • 当您准备重绘输出时,请写入db 0x1b,“[nA”向上移动n行。 (你需要把号码放在那里。)

  • Immediately after that (or after any subsequent newline), write db 0x1b, "[K" to clear that line. (You already knew this, but I'm including it for completeness.)
  • 紧接着(或任何后续换行符之后),写入db 0x1b,“[K”清除该行。 (你已经知道了这一点,但我将其包括在内是为了完整性。)

I wrote an example program to implement this, based partially on yours. It displays:

我写了一个示例程序来实现它,部分基于你的。它显示:

Hello, this is the clock program.
Line two.

Then, a short while later

不久之后

=== TEST ===
More.

And then

=== TEST 2 ===
Again.

This technique should be generalizable to any reasonable number of lines.

该技术应该可以推广到任何合理数量的线。

BITS 32
section .text

welcome      db "Hello, this is the clock program", 10, "Line two.", 10
welcome_len  equ $-welcome

test_str     db 0x1b, "[2A", 0x1b, "[K=== TEST ===", 10, 0x1b, "[KMore.", 10
test_len     equ $-test_str

test2_str     db 0x1b, "[2A", 0x1b, "[K=== TEST 2 ===", 10, 0x1b, "[KAgain.", 10
test2_len     equ $-test2_str

func_print:
    mov eax, 4
    mov ebx, 1
    int 0x80
    ret

pause: ; Note: DON'T EVER USE THIS IN A REAL PROGRAM. This is not how you sleep properly.
    mov eax, 0
loop:
    inc eax
    cmp eax, 1000000000
    jl loop
    ret

global  _start
_start:
    mov ecx, welcome
    mov edx, welcome_len
    call func_print

    call pause

    mov ecx, test_str
    mov edx, test_len
    call func_print

    call pause

    mov ecx, test2_str
    mov edx, test2_len
    call func_print

    mov eax, 1
    mov ebx, 0
    int 0x80