On Tue, Jul 20, 2010 at 9:36 PM, Rob Landley <[email protected]> wrote:
> On Monday 19 July 2010 22:16:40 Denys Vlasenko wrote:
>> Ah, FEATURE_**VI**_ASK_TERMINAL. I thought you test it with shell...
>
> I tested both, the behavior was the same. 80x25.
>
> (I'm assuming that hitting "enter" a couple times re-detects the shell input
> line size, and I don't have to exit and re-start the shell.)
>
>> FEATURE_VI_ASK_TERMINAL case is similar:
>>
>> # if ENABLE_FEATURE_VI_ASK_TERMINAL
>> if (!G.get_rowcol_error)
>> G.get_rowcol_error =
>> # endif
>> get_terminal_width_height(STDIN_FILENO, &columns,
>> &rows);
>>
>> I need to know whether here G.get_rowcol_error gets set to 1,
>> and if not, why... can you take a look?
>
> gre=1
>
> Yet when I resize the window and then try to cursor around in a file that goes
> off the right edge of the screen, it's very unhappy. (Well, if I make it
> bigger it happily gives me an 80x25 chunk in the upper left corner, and
> scrolls within it, but if I make it smaller it goes all wrappy and strange.)
vi detects window size only when it loads the file (see edit_file() function),
not every second. (It also does it on SIGWINCH, which you'll never get
on serial console).
Aha. I see the bug.
query_screen_dimensions();
#if ENABLE_FEATURE_VI_ASK_TERMINAL
if (G.get_rowcol_error /* TODO? && no input on stdin */) {
uint64_t k;
write1("\033[999;999H" "\033[6n");
fflush_all();
k = read_key(STDIN_FILENO, readbuffer, /*timeout_ms:*/ 100);
if ((int32_t)k == KEYCODE_CURSOR_POS) {
uint32_t rc = (k >> 32);
columns = (rc & 0x7fff);
rows = ((rc >> 16) & 0x7fff);
}
query_screen_dimensions(); <===== WRONG, DELETE THIS
}
#endif
The idea was to reuse sanitization code in query_screen_dimensions(),
but I overlooked that it will revert to 80x24.
Please try this patch:
http://busybox.net/downloads/fixes-1.17.0/busybox-1.17.0-vi.patch
--
vda
_______________________________________________
busybox mailing list
[email protected]
http://lists.busybox.net/mailman/listinfo/busybox
