Perhaps you fetched a new version from the ITRC; saved in on a Windows platform and FTP'd it in binary mode to a Unix server. In that case, you have linefeed+carriage-return characters instead of simple linefeeds as your line delimiters.
If you do 'cat -etv file' you will see:
^M$
at the line's termination.
If this is the case, eliminate the carriage returns with:
OK David. Fess up some more. This ain't HP-UX is it? My psychic, Miss Cleo, tells me that you are one of them Linux fellers. Why ain't you posting this in Linux where it belongs?
In any event, caljd.pl will work in Linux (or Windows for that matter) and that is what I consider to be the portable caljd.xx version.
Most bash implementations fail with this construct:
echo "Stupid Bash" | read A B echo "A = \"${A}\"" echo "B = \"${B}\""
${B} will be NULL rather than "Bash". You bash boys can argue about lvalues and subprocesses but this is a handy feature for parsing data.
Bash also does not know about "Z" (Zero-padded) variables.
To add: In fact, I submit that Clay should STOP parallel enhancements to caljd.sh and caljd.pl and only provide NEW features in caljd.pl [ assuming that anything is left of the alphabet :-) ]. In that fashion, this my-shell-ain't-your-shell nonsense would cease!
Well David, Jim has given me something to think about.
I will say that under Linux zsh is the least evil shell that I have found. Caljd.sh should work under zsh with one change: there is a line inside the AWK() function, "readonly AWKEXE". Comment out that line and zsh should work. Linux has a number of other 'improvements' such as the echo command. By default, echo "Test \c" outputs "Test \c" whereas in real UNIX, echo "Test \c" outputs "Test ". It would be easy enough to replace -Z variables with a function that pads data and I could probably add something that puts echo in a compatible mode and conditionally handle the echo "Stupid Bash" | read A B problem ("feature" if you are a bash guy) but Perl sounds like a whole lot easier answer.
After all, you don't have to know anything about Perl to use caljd.pl, just replace ".sh" in your scripts which call caljd.sh with ".pl" and declare victory.
One thing I will say with absolute confidence is that no version of caljd.sh that I write will ever run under csh.
OK David. Here is caljd.sh, Vrsn 2.3 s. Other than bug fixes, (which henceforth will not include "this doesn't work under my shell"), any newer features will be found only in the Perl version of the script.
In any event, if a stupid shell is detected which does not correctly (and I get to define 'correctly') handle echo "Stupid Bash" | read A B then I execute echo "Stupid Bash" > tempfile read A B < tempfile instead.
Similarly, if a stupid shell is detected which does not execute echo "Stupid Bash\c" correctly, I assert a shell option which changes the behavior of echo.
I also removed all references to "typeset -Z" because not all shells support -Z (zero-padded) variables. Instead, I call a function to do the padding.
I had already addressed a problem under Solaris. For some reason, Sun missed the memo about 20 years ago that nawk (new awk) could be renamed or linked to awk. For anyone wanting to use the original awk, oawk (old awk) is still available. In any event, if nawk is in the PATH, it will be used otherwise awk will be used.
Fine, David. You don't know how relieved I am to know that caljd.sh works under bash.
I suppose I should add that I did add a few comments about how the "shbang" line should be set so that the only change needed for various platforms should be changing the shbang line BUT I consider caljd.pl to be the much more portable answer because the ONLY change that might be needed is a symbolic link from wherever Perl really is to /usr/bin/perl.
