fixed a typo on the usage message, redone the test files that deal with with
Description:
Solution:
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
changes of input/ output format
Description:
on web page /RFC/h5diff
Solution:
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
changed algorythm
Description:
changed the array_diff function to increment the read buffer with a size selected
by an appropriate native data type.
previously it was assumed, e.g. int size = 4 bytes
Solution:
Platforms tested:
Linux 2.4 (rockaway)
SunOS 5.7 (arabica, 64bit mode)
IRIX 6.5 (modi4)
Misc. update:
bug fix
Description:
datasets with different storage datatypes where not read to memory to same size types
Solution:
obtain the memory type and size, for the TWO datasets ,
with the "fixtype" function
"upgrade" the smaller size to the larger if they are different
added H5DIFF_DEBUG flag for print information
Platforms tested:
Linux 2.4 (rockaway)
SunOS 5.7 (arabica, 64bit mode)
IRIX 6.5 (modi4)
Misc. update:
added new test files for h5diff
bug fix
Description:
the criteria for determining the correct switch case cast to void* to type* was the data size read from disk.
this was causing problems on Cray T3E, where all integer except char are 8 bytes long
Solution:
changed the criteria for using the memory size instead, which is determined by the same function
that reads the data into memory
the messages " using memory NATIVE_TYPE_X" were removed, because they might not coincid on different
platforms.
Platforms tested:
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
New feature/enhancement
Description:
Chunked datasets are handled poorly in several circumstances involving
certain selections and chunks that are too large for the chunk cache and/or
chunks with filters, causing the chunk to be read from disk multiple times.
Solution:
Rearrange raw data I/O infrastructure to handle chunked datasets in a much
more friendly way by creating a selection in memory and on disk for each chunk
in a chunked dataset and performing all of the I/O on that chunk at one time.
There are still some scalability (the current code attempts to
create a selection for all the chunks in the dataset, instead of just the
chunks that are accessed, requiring portions of the istore.c and fillval.c
tests to be commented out) and performance issues, but checking this in will
allow the changes to be tested by a much wider audience while I address the
remaining issues.
Platforms tested:
h5committested, FreeBSD 4.8 (sleipnir) serial & parallel, Linux 2.4 (eirene)
h5diff update
Description:
removed the -l , list, option
this feature will be part of a different tool
Solution:
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
new feature
Description:
during the sequencial match list, by default the differences between matched datasets were printed.
added an -m option that allows to turn this off, just printing the file differences
Solution:
Platforms tested:
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
added more tests, described in the test matrix
Description:
test -r
test -l
expanded -d, -p and -n tests
test different types to include links and named types
test for non supported classes, e.g bitfield, opaque , etc
Platforms tested:
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
added some niceties and utilities, and more tests
Description:
some niceties: more error messages on cases of bad input
utilities: some functions to more human readable output
more tests: described in the test matrix
Platforms tested:
Linux/rockaway(C)
SunOS/arabica (C)
SGI/modi4 (C)
added more tests to automatic testing
Description:
added test for an invalid option letter
added test for -h option
added test for an invalid option -d
added test for an invalid option -d number
Solution:
Platforms tested:
linux, solaris, sgi
Misc. update:
added test script for h5diff
code cleaning for alpha release
makefile now generates the h5difftst.c program that generates 2 .h5 files for testing
Description:
the .sh script runs several runs of h5diff and compares the output
with a predifined output located in /tools/testfiles (.txt files)
righ now it has only one test
Solution:
Platforms tested:
linux (other platforms later , ok !? )
Misc. update:
h5diff bug fix, enhancements
Description:
the case for 8 byte integers was not handled
added more checking of wrong input data
Solution:
added the size 8 integer case
Platforms tested:
w2000, linux, solaris (64 mode)
bug fix
Description:
some data not read correctly due to type_id used in read
Solution:
did a function that determines the optimal memory type
Platforms tested:
w2000, linux, solaris
bug fixes in h5diff
Description:
there was a incorrect reading of arguments when only 1 object name was supplied
reading of data was done using type_id got from disk, causing some values not read correctly into memory
Solution:
used native types to read into memory
Platforms tested:
w2000, linux, solaris57
h5diff enhancements
Description:
changed -r to count the number of differences, but do not print them
changed the output format with some printf enginnering, to align columns
print the dimensions of dsets, in case they are different
test files added test for non supported data (class not integer or float)
fix print dimensions to 0 index
Platforms tested:
w2000, linux, solaris
h5diff bugs and enhancements
Description:
added checking for not comparable datatypes
added checking for the same dimensionality
changed the functionality of -n option, to check up to n reported differences
changed the formula for -p to relative error with absolute value
fixed a bug on solaris, related to printf integer format (int datatype)
Platforms tested:
w2000, linux, solaris
removed -d , -g , -t and -v options and renamed -m to -d and -v to -l
added output for objects that are not in one file but in the other
Platforms tested:
windows, linux
bug fix
added more verbose help message
added support for all dataset types
added more error checking
added new output format (28 Jan version )
Platforms tested:
windows , linux
