- 09 Jun, 2017 3 commits
-
-
Roland Haas authored
they do not have to be the same: https://www.mcs.anl.gov/research/projects/mpi/mpi-standard/mpi-report-2.0/node47.htm
-
Roland Haas authored
this is a bit clearer than hard-coding the relation between the request number and the worker it was sent to
-
this makes sure that only one rank uses O_TRUNC but I can still use fwrite() to get stdio buffering (for free)
-
- 14 Apr, 2017 1 commit
-
-
Roland Haas authored
this is mostly so that I don't confuse 'e' for exit and 'E' for error
-
- 13 Apr, 2017 1 commit
-
-
Roland Haas authored
-
- 15 Mar, 2017 18 commits
-
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
- 14 Mar, 2017 4 commits
-
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
namely for pax do not ever user ther prefix field
-
- 12 Mar, 2017 6 commits
-
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
instead look up format description in POSIX
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
-
- 11 Mar, 2017 7 commits
-
-
Roland Haas authored
-
Roland Haas authored
* run testsuite when test script changes * declare all target as phony
-
Roland Haas authored
I should also test for huge files, but given that huge files only occur at 8GB this would be an extensive test
-
Roland Haas authored
-
Roland Haas authored
-
Roland Haas authored
this is not really required in order for the file being a correct pax file but it makes comparison easier
-
Roland Haas authored
apparently the last \0 is optional if the file name would otherwise overflow the field?
-