Marc Lehmann --> rxvt-unicode (2009-04-09 05:14:25 +0200):
[...]
> Maybe you would be able to reproduce this under ktrace (on bash), so one
> can see whether bash really writes the prompt or indeed bash just doesn't
> write the prompt.
Sure, see below. Bash receives SIGWINCH after having written the first
of the two prompt lines. It then determines the screen size and resumes
writing the prompt. However, before _and after_ writing the second line
of the prompt, the line is cleared (and the second sequence even followed
by \c\b\b), thus the second line doesn't show up...
[...]
12232 1 bash CALL write(2,0x80fc808,0x1a)
12232 1 bash GIO fd 2 wrote 26 bytes
"ju...@moray:~ [17:37:43]\r\n"
12232 1 bash RET write 26/0x1a
12232 1 bash PSIG SIGWINCH caught handler=0x80cb014 mask=(2):
code=SI_USER sent by pid=14142, uid=1000)
12232 1 bash CALL ioctl(0,TIOCGWINSZ,0xbfbfd85c)
12232 1 bash GIO fd 0 read 8 bytes
"T\0b\0\^P\^C\M^X\^D"
12232 1 bash RET ioctl 0
12232 1 bash CALL write(2,0x80fc808,6)
12232 1 bash GIO fd 2 wrote 6 bytes
"\r\^[[K$ "
12232 1 bash RET write 6
12232 1 bash CALL __sigprocmask14(3,0x80fae5c,0)
12232 1 bash RET __sigprocmask14 0
12232 1 bash CALL setcontext(0xbfbfd934)
12232 1 bash RET ptrace JUSTRETURN
12232 1 bash CALL write(2,0x80fc808,7)
12232 1 bash GIO fd 2 wrote 7 bytes
"\r\^[[K\r\b\b"
12232 1 bash RET write 7
12232 1 bash CALL __sigprocmask14(0,0,0x80fbc78)
12232 1 bash RET __sigprocmask14 0
12232 1 bash CALL read(0,0xbfbfdddf,1)
12232 1 bash GIO fd 0 read 1 bytes
"#"
12232 1 bash RET read 1
[...]
(The last read(2) shown above successfully reads the first character (#)
of the command I entered after the prompt was (incorrectly) printed.)
Moving to bug-bash now... ;-)
Regards, Jukka
--
This email fills a much-needed gap in the archives.
_______________________________________________
rxvt-unicode mailing list
[email protected]
http://lists.schmorp.de/cgi-bin/mailman/listinfo/rxvt-unicode
