Changeset 1075 in MondoRescue for branches/stable/mondo/src/common/libmondo-fifo.c

Timestamp:

Jan 25, 2007, 4:18:59 PM (17 years ago)

Author:

Bruno Cornec

Message:

Improvements for USB support in mindi when called from mondo
Changelogs in C files removed from common

File:

• branches/stable/mondo/src/common/libmondo-fifo.c (modified) (1 diff)

branches/stable/mondo/src/common/libmondo-fifo.c

@@ -1 +1 @@
 /* libmondo-fifo.c
    $Id$
-
-
-04/17
-- replaced INTERNAL_TAPE_BLK_SIZE with bkpinfo->internal_tape_block_size
-
-04/04/2004
-- ps wax is now ps wwax
-
-09/10/2003
-- p75 is now p80 in outopening tape call
-- negotiate largest buffer possible when opening tape drive
-- fall back to dd if buffer fails
-
-09/09
-- better logging for tape users
-- working on try_hard_to_fwrite(), try_hard_to_fread()
-- replaced internal w/ EXTERNAL 'buffer' exe.
-
-08/30
-- tweaked error msgs in try_hard_to_fwrite()
-
-08/02
-- updated is_incoming_block_valid() to make it
-  return end-of-tape if >300 flotsam blocks
-
-05/02
-- when logging tape errors, don't repeat self
-
-04/24
-- added lots of log_OS_error()'s and assert()'s
-
-04/22
-- copy_file_from_here_to_there() --- added a bit of fault tolerance;
-  if write fails, retry a few times before reporting error
-
-04/07/2003
-- line 866 --- set block_size used by internal buffer to 32768;
-  was INTERNAL_TAPE_BLK_SIZE/2
-
-10/01 - 11/30/2002
-- is_incoming_block_valid() --- always make
-  checksums %65536, just in case int size is
-  odd (GRRR, ArkLinux)
-- disabled rotor-related superfluous warnings
-- added INTERNAL_TAPE_BLK_SIZE
-- do_sem() now returns int explicitly
-- changed internal_block_size
-- some irregularities (e.g. bad 'type'-ing) found by Kylix; fixed by Hugo
-
-09/01 - 09/30
-- change 64k to TAPE_BLOCK_SIZE
-- added internal_block_size; set it to TAPE_BLOCK_SIZE*2
-- if data is flowing FROM tape TO hard disk then set the threshold to 10 (not 75)
-- lots of multitape-related fixes
-- finally caught & fixed the 'won't finish unzipping last bigfile' bug
-- added WIFEXITED() after waitpid(), to improve multi-tape support
-
-08/01 - 08/31
-- trying to catch & stop 'won't finish unzipping last bigfile' bug by
-  changing the copy_file_rom_here_to_there code
-- changed % threshold from 95 back to 75
-- don't insist on full 256K write of last block to tape
-- if >25 secs go by & all data (AFAIK) has been copied thru by FIFO wrapper
-  and g_tape_stream is _still_ not closed then shrug shoulders & pthread_exit
-  anyway...
-- change fprintf()'s to log_it()'s
-- added a header+footer to each block as it is read/written to/from tape
-  by copy_file_from_here_to_there
-- wrote workaround to allow >2GB of archives w/buffering
-- changed % threshold from 75 to 95
-- added calls to set_signals(); 'buffer' was killing mondoarchive as
-  it terminated
-- cleaned up struct-passing, to improve reliability and eliminate
-  some race conditions
-- changed some forks to pthreads
-- added some comments
-- created libfifo{.c,.h,-EXT.h}
-- copied core of 'buffer' here
-- added some other, Mondo-specific functions
-- hacked 'buffer' into user-friendliness
 */