Changes between Initial Version and Version 1 of Ticket #17867, comment 4


Ignore:
Timestamp:
Sep 9, 2022, 10:15:39 PM (21 months ago)
Author:
bipolar

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #17867, comment 4

    initial v1  
    33At least on Linux, tcdrain() waits until all the bytes in the output buffer are actually written.
    44
    5 That's doesn't seems to be the case on Haiku in my tests, and instead of relying on tcdrain() blocking, I have to call `sleep()` with a number of milliseconds proportional to the baudrate I'm using, to make sure the data is properly sent.
     5That doesn't seems to be the case on Haiku in my tests, and instead of relying on tcdrain() blocking, I have to call `sleep()` with a number of milliseconds proportional to the baudrate I'm using, to make sure the data is properly sent.
    66
    77Double checking due to your comment, it does indeed like Haiku's tty module takes that TCSBRK and TTYSETBREAK to usb_serial (a no-op there) and to pc_serial drivers (where it seems to set/unset a register, but not block till the output buffer is emptied).
     
    3535Writing this just for reference, or to let others spots where I'm getting confused.
    3636
    37 In any case, I guess I should test some more (even under BeOS), write some test cases I can share, and hopefully understand enough as to, at least, propose some changes.
     37In any case, I guess I should test some more (even under BeOS), write some test programs I can share, and hopefully understand enough as to, at least, propose some changes.
    3838
    3939Pointers, suggestions and corrections, always welcomed! Thanks!