Eric Blake wrote:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
According to Larry Breyer on 9/27/2006 3:22 PM:
What changed from bash 3.1.17 to 3.1.18 ?
I don't know, but let me know when you find out :)
(There is no 3.1.18 - the latest official bash is 3.1.17, at cygwin
release 8, but upstream is very unlikely to release 3.1.18 since 3.2 is in
beta now.)
I am a bit confused by these numbering schemes. When I look at the
'Select Packages' section of setup, my choices for bash are 3.1-6 or
3.1-8. I assumed they were shorthand for 3.1.16 and 3.1.18.
I got syntax errors on blank lines.
That means your script has CRLF line endings, but resides on a binary
mount point.
I think I panicked slightly at discovering Cygwin bash had been
regressed. I was thinking of all the supplied scripts so I ran
a test of the ~450 shell scripts and found only a handful with CR
embedded:
/bin/card
/bin/fixps
/bin/flea
/bin/makempx
/bin/mktexlsr
/bin/nfs-server-config
/etc/postinstall/uw-imap-imapd.sh.don
/etc/postinstall/xerces-c-devel.sh.done
/etc/preremove/gv.sh
/etc/preremove/uw-imap-imapd.sh
/etc/preremove/X-start-menu-icons.sh.done
/etc/preremove/xerces-c-devel.sh
I found several of the scripts I wrote had embedded CRs. I'm an
old Unix hand, so I am really baffled how they got into my scripts.
Plus, I always found vim would display the CRs as ^M. To my dismay
I now discover the vim silently ignores CR when its part of CR-LF
pair, unlike LF-CR.
So, I have subscribed to cygwin and cygwin-announce. I am hopeful I
will not be surprised again.
--
Larry Breyer
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
