I'm trying to configure a diff tool in TortoiseHg and Mercurial.
This is the configuratión I've done:
[merge-tools]
CmpIt.executable = E:\Program Files (x86)\Compare It!\wincmp3.exe
CmpIt.priority = 1
CmpIt.premerge = True
CmpIt.gui = True
CmpIt.args= -m $local $base $other
CmpIt.diffargs= -m $local $base
CmpIt.dirdiff= True
But when the tool start, it's looking for "$local" and "$base" files instead real file names.
I found on Task Manager that the command line is geting the names of the args instead their values:
E:\Program Files (x86)\Compare It!\wincmp3.exe -m $local $base
How can I force to pass the value of the args to the command line?
You miss some important points
You must to avoid usage CompareIt! as merge-tool (it still doesn't support 3-way merge): CmpIt.args is BIG NO-NO for now
You have to re-read hgrc.d\MergeTools.rc (of any merge-tool part) for parameters, used in .diffargs, parred by TortoiseHG/Mercurial: $local + $base are variables for merge, not for diff
You have to re-read CompareIt! command-line options: they doesn't mention -m at all
From my POV, using bcompare.diffargs options as quick-start and CompareIT! Help for "translation", this will reasonably good first iteration (not tested!!!)
CmpIt.args= -m $local $base $other
CmpIt.diffargs= $parent /=$plabel1 $child /=$clabel
Related
Same source file.
directe compile use valac.
⭕ valac --pkg gtk+-3.0 -X -lm --pkg libcanberra src/Application.vala
⭕ ls Application
-rwxrwxr-x 1 eexpss 48K 05-13 19:59 Application
here is part of my meson.build.
project('com.github.eexpress.cairo-timer', 'vala', 'c')
# i18n = import('i18n')
executable(
meson.project_name(),
'src/Application.vala',
dependencies: [
dependency('gtk+-3.0'),
# dependency('cairo'),
dependency('libcanberra')
],
# link_args : '-X',
# link_args : '-lm',
link_args : ['-X', '-lm',],
install: true
)
and use ninja to compile it.
⭕ cd build; ninja
⭕ ls com.github.eexpress.cairo-timer
-rwxrwxr-x 1 eexpss 98K 05-13 17:02 com.github.eexpress.cairo-timer
So the binary file is more larger than above one. why?
Because you didn't enable debugging for valac, but meson enables it by default. Add -g to valac and the output size should be close to equal.
To see how ninja and valac run the tools to build, enable verbose option by given -v to both commands.
The minor size differences are, as I assume, from file names in them. Compare the outputs, for example, from readelf --debug-dump=line hello to see the diff.
I'm trying to checkout Joe English's tile-extras from Github using svn using a Tcl script.
The required command is
svn checkout https://github.com/jenglish/tile-extras.git path
I have some code that boils down to
exec C:/cygwin64/bin/svn.exe checkout \
https://github.com/jenglish/tile-extras.git C:/cygwin64/tmp/TCL61416]
which fails with the message
couldn't execute "C:\cygwin64\bin\svn.exe checkout
https:\github.com\jenglish\tile-extras.git
C:\cygwin64\tmp\TCL61416": No error
Pasting the command quoted in the error message into a Windows Command Prompt window, I see
svn: E125002: 'https:\github.com\jenglish\tile-extras.git' does not appear to be a URL
So, the problem seems to be that exec converts Tcl-style paths to Unix-style a little over-enthusiastically. Is there any way I can prevent it from converting https://github.com/jenglish... to https:\github.com\jenglish...?
For information, I'm running on Windows 10, with cygwin (setup version 2.889 (64 bit)), svn 1.9.7 and tcl version 8.6.7 (via ActiveTcl 8.6.7.0).
UPDATE
Here is my actual code, which I'm only slightly embarrassed by:
# svn wrapper proposed by Donal Fellows at
# http://stackoverflow/questions/49224268
proc svn {args} {
exec {*}[auto_execok svn] {*}$args <#stdin >#stdout }
# Checkout from github to a temporary repository
set repository https://github.com/jenglish/tile-extras.git set
svnImage [auto_execok svn]
set fil [file tempfile tempfnm] close $fil file delete $tempfnm
set tempRepo [file rootname $tempfnm] puts stdout tempRepo:\ $tempRepo
file mkdir $tempRepo
set svnCmd [list svn checkout $repository [file nativename $tempRepo]]
puts stdout svnCmd:\ $svnCmd eval $svnCmd
# Determine the tile-extras sources
set sourceFiles {keynav.tcl icons.tcl}
set targets [file nativename [file join $tempRepo trunk *.tcl]]
foreach filnam [split [svn ls $targets] \n] {
if {[string match *.tcl $filnam] && [lsearch $sourceFiles $filnam] < 0} {
lappend sourceFiles $filnam
}
}
And here is the result
$ tclsh foo.tcl
tempRepo: C:/cygwin64/tmp/TCL61838
svnCmd: svn checkout
https://github.com/jenglish/tile-extras.git {C:\cygwin64\tmp\TCL61838}
A C:\cygwin64\tmp\TCL61838/branches
A C:\cygwin64\tmp\TCL61838/trunk
A C:\cygwin64\tmp\TCL61838/trunk/README.md
A C:\cygwin64\tmp\TCL61838/trunk/dialog.tcl
A C:\cygwin64\tmp\TCL61838/trunk/doc
A C:\cygwin64\tmp\TCL61838/trunk/doc/dialog.n
A C:\cygwin64\tmp\TCL61838/trunk/doc/keynav.n
A C:\cygwin64\tmp\TCL61838/trunk/icons.tcl
A C:\cygwin64\tmp\TCL61838/trunk/keynav.tcl
A C:\cygwin64\tmp\TCL61838/trunk/license.terms
A C:\cygwin64\tmp\TCL61838/trunk/pkgIndex.tcl
Checked out revision 7.
svn: E155007: '/home/alan/C:\cygwin64\tmp\TCL61838\trunk\*.tcl' is not a working copy
while executing "exec {*}[auto_execok svn] {*}$args <#stdin >#stdout"
(procedure "svn" line 2)
invoked from within "svn ls $targets"
invoked from within "split [svn ls $targets] \n"
invoked from within "foreach filnam [split [svn ls $targets] \n] {
if {[string match *.tcl $filnam] && [lsearch $sourceFiles $filnam] < 0} {
lappend sourceFiles $filn..."
(file "foo.tcl" line 30)
$ ls /tmp/TCL61838/
$
The directory /tmp/TCL61838 is empty, so it seems the svn checkout command didn't complete completely happily. I also see an unpleasant mixture of forward slashes and backslashes being reported by svn.
Thanks in advance for any more help.
Given the error message, it looks like you're getting word boundaries wrong in the code that you've not shown us; while you might believe the code “boils down to” to that exec, it's not actually done that. Also, you've flipped the slashes in the URL which won't work, but that's probably a side-effect of something else.
Alas, I can't quite guess how to fix things for you. There's just too many options. I provide a suggestion below, but there's no telling for sure whether it will work out.
Diagnosis Methodology
The evidence for why I believe that the problem is what I say? This interactive session log (on OSX, but the generic behaviour should be the same):
% exec cat asdkfajh
cat: asdkfajh: No such file or directory
% exec "cat akjsdhfdkj"
couldn't execute "cat akjsdhfdkj": no such file or directory
% exec "cat aksdjhfkdf" skdjfghd
couldn't execute "cat aksdjhfkdf": no such file or directory
The first case shows an error from an external program. The second case shows an error due to no-such-program. The third case shows that arguments are not reported when erroring due to to no-such-program.
This lets me conclude that both C:\cygwin64\bin\svn.exe and its arguments (checkout, https:\github.com\jenglish\tile-extras.git and C:\cygwin64\tmp\TCL61416) were actually passed as a single argument to exec, a fairly common error, and that the problems lie in the preparatory code. You don't show us the preparatory code, so we can't truly fix things but we can make suggestions that address the common problems.
Suggested Approach
A good way to reduce these errors is to write a small wrapper procedure:
proc svn {args} {
# I add in the I/O redirections so svn can ask for a password
exec {*}[auto_execok svn] {*}$args <#stdin >#stdout
}
This would let you write your call to svn as:
svn checkout $theURL [file nativename $theDirectory]
and it would probably Just Work™. Also note that only the directory name goes through file nativename; the URL does not. (We could embed the call to file nativename in the procedure if we were making a specialised procedure to do checkouts, but there's too much variation in the full svn program to let us do that. The caller — you — has to deal with it.)
I used to run a tcl script for the Cadence tools on a server, however, now the script fails to run.
The script is based on the following one:
#### Template Script for RTL->Gate-Level Flow
#### all basic steps except for DFT-scan
#### Fill in the <...> fields for your module
#### and update library search paths for your system
if {[file exists /proc/cpuinfo]} {
sh grep "model name" /proc/cpuinfo
sh grep "cpu MHz" /proc/cpuinfo
}
#### Set up
set DESIGN test
set SYN_EFF medium
set MAP_EFF medium
set DATE test
set global_map_report 1
set map_fancy_names 1
set iopt_stats 1
set SYN_PATH "."
set _OUTPUTS_PATH outputs_${DATE}
set _LOG_PATH logs_${DATE}
set _REPORTS_PATH reports_${DATE}
set_attribute lib_search_path {. ./lib} /
set_attribute hdl_search_path {. ./rtl} /
set_attribute information_level 7 /
set_attribute map_timing true /
set_attribute retime_reg_naming_suffix __retimed_reg /
set_attribute library lib
... continues
First I open a csh in order to run a csh script to setup the Cadence tools on the server, then I run source script.tcl. This used to work, however, now it fails with the following error:
Missing ].
And if I comment the first if:
set: Syntax Error.
What may have changed in the server for this to happen and how can I fix this? The script did not change, so its syntax is correct.
As your comments in code described, you should call source script.tcl inside the tool, not in csh which doesn't understand Tcl syntax, nor in tclsh which doesn't understand those Cadence-specific Tcl commands.
Also, the two line
sh grep "model name" /proc/cpuinfo
sh grep "cpu MHz" /proc/cpuinfo
should be
exec grep "model name" /proc/cpuinfo
exec grep "cpu MHz" /proc/cpuinfo
since exec is the correct Tcl command for calling shell commands.
I started shell scripting for my work, but I must admit I'm still far away from even being a rookie. Therefore I wanted to ask you for your help/advice.
I build a script for a big data application (taking the quick and dirty approach, patching stuff from the internet together) to recursively go through a folder structure and convert convert all XML files to JSON.
The status quo of my script is:
#!/bin/sh
# Shell script to find out all the files under a directory and
#its subdirectories. This also takes into consideration those files
#or directories which have spaces or newlines in their names
cd /Users/q337498/Desktop/Archiv/2014/01/10
DIR="."
function list_files()
{
if !(test -d "$1")
then echo $1; return;
fi
cd "$1"
#echo; echo `pwd`:; #Display Directory name
for i in *
do
if test -d "$i"; then # if dictionary
if [ "$(ls -A $i)" ]; then
list_files "$i" #recursively list files
cd ..
else
echo "$i is Empty"
fi
else
java -jar /Users/q337498/Desktop/XML2JSON/SaxonEE9-5-1-4J/saxon9ee.jar -s:"$i" -xsl:/Users/q337498/Desktop/xsltjson-master/conf/xml-to-json.xsl -o:output/$(pwd)/${i%%[.]*}
# if jsonlint /Users/q337498/Desktop/Archiv/2014/01/08/$(pwd)/${i%%[.]*} -q; then
# echo "GOOD"
# else
# echo "NOT GOOD"
# fi
# echo ${i%%[.]*}
# echo "$i"; #Display File name
fi
done
}
if [ $# -eq 0 ]
then list_files .
exit 0
fi
for i in $*
do
DIR="$1"
list_files "$DIR"
shift 1 #To read next directory/file namedone
done
This code works, but the problem is that for 60000 files it takes up to 15 hours on a macbook pro with 16gb RAM and an 2.8 Ghz i7. And I need to convert 10 million files.
How do you think that I could accelerate the script? parallelize? take some commands out? What options do I have, and how would I actually implement them?
The files are ultimately going to end up in a MongoDB, so if someone knows a better way to convert xml to json and upload it to mongo his input is also welcome.
Cheers,
Dudu
I see 2 immediate problems here:
You are invoking java once for each file, therefore incurring the JVM startup time for every file which is going to add up to a huge chunk of time.
You are running single threaded
So I would suggest that you:
Write a Java program that does the directory traversal and does your transformation
Benchmark the performance difference
Try other Java libraries for doing the XML->JSON conversion https://github.com/beckchr/staxon/wiki/Benchmark
If necessary for performance, add multi-threading to your application using java.util.concurrent.
I had a quick look on the forums and I don't think this question has been asked already.
I am currently working with an MPI/CUDA hybrid code, made by somebody else during his PhD.
Each CPU has its own GPU.
My task is to gather data by running the (already working) code, and implement extra things.
Turning this code into a single CPU / Multi-GPU one is not an option at the moment (later, possibly.).
I would like to make use of performance profiling tools to analyse the whole thing.
For now an idea is to have each CPU launch nvvp for its own GPU and gather data, while another profiling tool will take care of general CPU/MPI part (I plan to use TAU, as I usually do).
Problem is, launching nvvp's interface 8 simultaneous times (if running with 8 CPU/GPUs) is extremely annoying. I would like to avoid going through the interface, and get a command line that directly writes the data in a file, that I can feed to nvvc's interface later and analyse.
I'd like to get a command line that will be executed by each CPU and will produce for each of them a file giving data about their own GPU. 8 (GPUs/CPUs) = 8 files.
Then I plan to individually feed and analyse these files with nvcc one by one, comparing the data manually.
Any idea ?
Thanks !
Take a look at nvprof, part of the CUDA 5.0 Toolkit (currently available as a release candidate). There are some limitations - it can only collect a limited number of counters in a given pass and it cannot collect metrics (so for now you'd have to script multiple launches if you want more than a few events). You can get more information from the nvvp built-in help, including an example MPI launch script (copied here but I suggest you check out the nvvp help for an up-to-date version if you have anything newer than the 5.0 RC).
#!/bin/sh
#
# Script to launch nvprof on an MPI process. This script will
# create unique output file names based on the rank of the
# process. Examples:
# mpirun -np 4 nvprof-script a.out
# mpirun -np 4 nvprof-script -o outfile a.out
# mpirun -np 4 nvprof-script test/a.out -g -j
# In the case you want to pass a -o or -h flag to the a.out, you
# can do this.
# mpirun -np 4 nvprof-script -c a.out -h -o
# You can also pass in arguments to nvprof
# mpirun -np 4 nvprof-script --print-api-trace a.out
#
usage () {
echo "nvprof-script [nvprof options] [-h] [-o outfile] a.out [a.out options]";
echo "or"
echo "nvprof-script [nvprof options] [-h] [-o outfile] -c a.out [a.out options]";
}
nvprof_args=""
while [ $# -gt 0 ];
do
case "$1" in
(-o) shift; outfile="$1";;
(-c) shift; break;;
(-h) usage; exit 1;;
(*) nvprof_args="$nvprof_args $1";;
esac
shift
done
# If user did not provide output filename then create one
if [ -z $outfile ] ; then
outfile=`basename $1`.nvprof-out
fi
# Find the rank of the process from the MPI rank environment variable
# to ensure unique output filenames. The script handles Open MPI
# and MVAPICH. If your implementation is different, you will need to
# make a change here.
# Open MPI
if [ ! -z ${OMPI_COMM_WORLD_RANK} ] ; then
rank=${OMPI_COMM_WORLD_RANK}
fi
# MVAPICH
if [ ! -z ${MV2_COMM_WORLD_RANK} ] ; then
rank=${MV2_COMM_WORLD_RANK}
fi
# Set the nvprof command and arguments.
NVPROF="nvprof --output-profile $outfile.$rank $nvprof_args"
exec $NVPROF $*
# If you want to limit which ranks get profiled, do something like
# this. You have to use the -c switch to get the right behavior.
# mpirun -np 2 nvprof-script --print-api-trace -c a.out -q
# if [ $rank -le 0 ]; then
# exec $NVPROF $*
# else
# exec $*
# fi
Another option is since you are already using TAU to profile the CPU side of the application you could also use TAU to collect the GPU performance data. TAU supports multi-gpu execution along with MPI, take a look at http://www.nic.uoregon.edu/tau-wiki/Guide:TAUGPU for instructions on how to get started using TAU's GPU profiling capabilites. TAU uses CUPTI (CUda Performance Tools Interface) underneath and so the data you will be able to collect with TAU will be very similar to what to can collect with nVidia's Visual Profiler.
Things have changed since CUDA 5.0 and now we can simply use %h, %p and %q{ENV} as mentioned here instead of using a wrapper script:
$ mpirun -np 2 -host c0-0,c0-1 nvprof -o output.%h.%p.%q{OMPI_COMM_WORLD_RANK} ./my_mpi_app
Apparently since 2015 it is possible to auto-annotated MPI calls via NVTX and mpi_interceptions.so library when using nvprof profiler:
https://devblogs.nvidia.com/gpu-pro-tip-track-mpi-calls-nvidia-visual-profiler/
http://on-demand.gputechconf.com/gtc/2017/presentation/s7495-jain-optimizing-application-performance-cuda-profiling-tools.pdf
TAO still does not support distributed deep learning according to this presentation:
http://on-demand.gputechconf.com/gtc/2017/presentation/s7684-allen-malony-performance-analysis-of-cuda-deep-learning-networks-using-tau.pdf