I think bash logic in a loop like : while [ condition] |read somevar is flawed.
Subject: [50 character or so descriptive subject here (for reference)] Configuration Information [Automatically generated, do not change]: Machine: x86_64 OS: linux-gnu Compiler: gcc -I/home/abuild/rpmbuild/BUILD/bash-4.2 -L/home/abuild/rpmbuild/BUILD/bash-4.2/../readline-6.2 Compilation CFLAGS: -DPROGRAM='bash' -DCONF_HOSTTYPE='x86_64' -DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='x86_64-suse-linux-gnu' -DCONF_VENDOR='suse' -DLOCALEDIR='/usr/share/locale' -DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H -I. -I. -I./include -I./lib -fmessage-length=0 -O2 -Wall -D_FORTIFY_SOURCE=2 -fstack-protector -funwind-tables -fasynchronous-unwind-tables -g -D_GNU_SOURCE -DRECYCLES_PIDS -Wall -g -std=gnu89 -Wuninitialized -Wextra -Wno-unprototyped-calls -Wno-switch-enum -Wno-unused-variable -Wno-unused-parameter -ftree-loop-linear -pipe -fprofile-use uname output: Linux dep2.fritz.box 3.7.10-1.16-default #1 SMP Fri May 31 20:21:23 UTC 2013 (97c14ba) x86_64 x86_64 x86_64 GNU/Linux Machine Type: x86_64-suse-linux-gnu Bash Version: 4.2 Patch Level: 42 Release Status: release bug-bash@gnu.org Description: [Detailed description of the problem, suggestion, or complaint.] Hello, this script: ___ export cval=0 echo entering while # listing ten files, to generate some output to count. ls -1 /usr/bin|head -10 |while read fname do cval=$(( cval +1 )) echo cval = $cval file = $fname done # one would think cval is now 10. but it is not, contrary to any other programming language echo " after while: cval = $cval" ___ does not set the value of cval after exiting the while loop. that makes no sense. Please Don't bother to tell me it s the way you guys think it should technically work. I know you think something along those lines, at least thats what I am told However, no one in this world having to solve real life issues with software, is interested in how it technically works. we need real life logic and software that can deal with real life challenges and requirements. Bash should not work that way. no programming language handles logic this way. Not pascal, korn shell, c shell, cobol, c, c++, lua , fortran or any other language i ever used. Thank you for your willingness to open up your trench line, if any. best regards, Rens Repeat-By: [Describe the sequence of events that causes the problem to occur.] Fix: [Description of how to fix the problem. If you don't know a fix for the problem, don't include this section.]
Re: I think bash logic in a loop like : while [ condition] |read somevar is flawed.
On 2013-12-23 23:57:32 +0100, rens wrote: > Hello, > > this script: > ___ > export cval=0 > echo entering while > > # listing ten files, to generate some output to count. > ls -1 /usr/bin|head -10 |while read fname Just use a `for` loop and a glob. Really. It's that easy. > do > cval=$(( cval +1 )) > echo cval = $cval file = $fname > done > # one would think cval is now 10. but it is not, contrary to any other > programming language Really, all other programming languages propagate variables from children to parents? I find that hard to believe. > that makes no sense. Yes, it does. You created a subshell. How do you expect for the variable to propagate back to the parent, where you are expanding it? > Please Don't bother to tell me it s the way you guys think it should > technically work. > I know you think something along those lines, at least thats what I am > told Don't use a subshell if you don't want one. > However, no one in this world having to solve real life issues with > software, is interested in how it technically works. we need real life > logic and software that can deal with real life challenges and requirements. Then stop using a subshell. > Bash should not work that way. no programming language handles logic this > way. Not pascal, korn shell, c shell, cobol, c, c++, lua , fortran or any > other language i ever used. None of the languages you listed propagate variables from children to parents directly. I don't know why you think you should have that variable in the parent when you populated it in a subshell... pgp6pXgFwKge0.pgp Description: PGP signature
Re: I think bash logic in a loop like : while [ condition] |read somevar is flawed.
On Tue, Dec 24, 2013 at 12:57 AM, rens wrote: > > Subject: [50 character or so descriptive subject here (for reference)] > > Configuration Information [Automatically generated, do not change]: > Machine: x86_64 > OS: linux-gnu > Compiler: gcc -I/home/abuild/rpmbuild/BUILD/bash-4.2 > -L/home/abuild/rpmbuild/BUILD/bash-4.2/../readline-6.2 > Compilation CFLAGS: -DPROGRAM='bash' -DCONF_HOSTTYPE='x86_64' > -DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='x86_64-suse-linux-gnu' > -DCONF_VENDOR='suse' -DLOCALEDIR='/usr/share/locale' -DPACKAGE='bash' > -DSHELL -DHAVE_CONFIG_H -I. -I. -I./include -I./lib -fmessage-length=0 > -O2 -Wall -D_FORTIFY_SOURCE=2 -fstack-protector -funwind-tables > -fasynchronous-unwind-tables -g -D_GNU_SOURCE -DRECYCLES_PIDS -Wall -g > -std=gnu89 -Wuninitialized -Wextra -Wno-unprototyped-calls -Wno-switch-enum > -Wno-unused-variable -Wno-unused-parameter -ftree-loop-linear -pipe > -fprofile-use > uname output: Linux dep2.fritz.box 3.7.10-1.16-default #1 SMP Fri May 31 > 20:21:23 UTC 2013 (97c14ba) x86_64 x86_64 x86_64 GNU/Linux > Machine Type: x86_64-suse-linux-gnu > > Bash Version: 4.2 > Patch Level: 42 > Release Status: release > bug-bash@gnu.org > Description: > [Detailed description of the problem, suggestion, or complaint.] > > Hello, > > this script: > ___ > export cval=0 > echo entering while > > # listing ten files, to generate some output to count. > ls -1 /usr/bin|head -10 |while read fname > do > cval=$(( cval +1 )) > echo cval = $cval file = $fname > done > # one would think cval is now 10. but it is not, contrary to any other > programming language > > echo " after while: cval = $cval" > ___ > > does not set the value of cval after exiting the while loop. > > that makes no sense. > > Please Don't bother to tell me it s the way you guys think it should > technically work. > I know you think something along those lines, at least thats what I am > told > > However, no one in this world having to solve real life issues with > software, is interested in how it technically works. we need real life > logic and software that can deal with real life challenges and requirements. > > Bash should not work that way. no programming language handles logic this > way. Not pascal, korn shell, c shell, cobol, c, c++, lua , fortran or any > other language i ever used. > Actually only ksh93 doesn't work this way, other "ksh" variants work like bash. In newer bash you can do, if job control is disable (that is in a script): shopt -s lastpipe. Note that even with this, only the right hand side doesn't run in a forked subshell, eg in ksh93 or bash with last pipe: var=foo | true;echo $var # prints nothing At least one side of the pipe has to be in a child process. The same thing is true if you use a pipe and fork a process in these other languages you mention.
Re: soft-bug (should be in planning queue)...
On 2013-12-21 14:13:25 -0800, Linda Walsh wrote: > Did you mean july 2013? Probably not, plenty of 4.3 code is not being backported. pgprkzu2zkgSh.pgp Description: PGP signature
