Mail Archives: cygwin/2011/08/25/13:46:29
On 8/25/2011 12:51 AM, Ronald Fischer wrote:
> On Wed, 24 Aug 2011 11:56 -0400, "Eliot Moss"<moss AT cs DOT umass DOT edu> wrote:
>> On 8/24/2011 11:52 AM, Ronald of Steiermark wrote:
>>> On Wed, 24 Aug 2011 08:39 -0700, "Andrew DeFaria"<Andrew AT DeFaria DOT com>
>>> wrote:
>>> For instance, to test the cruel BAT files which we are going to deliver.
>> You can run .bat files from bash and other Cygwin shells.
> You can *run* them, but the effect is not always the same.
>
> For example, setting an environment variable within a batch file under
> CMD.EXE results in the environment variable being visible in the calling
> environment (similar to "sourcing" a file in bash), while calling the
> batch file from bash leaves the environment intact.
This behavior is well known. However why does it matter? Are you saying
that you have to "source" a .bat file to set up the environment and then
execute another .bat file that requires those environment settings???
I'm not saying this doesn't happen but it's not a very good practice.
> Also, some internal commands (for example COPY) are not present in bash,
> though this can be easily remedied using an alias or a shell function.
Why should this be an issue? I mean if you exec some .bat file and it
calls "copy" does it not work?
> Other problems are related to the use of \ as a path separator. Imagine
> that some of your BATCH files generate environment variables containing
> a Windows path,
Again, why is this a problem? If your .bat file has "copy
C:\Windows\some.dll C:\Temp" does it not work? Or are you saying that
you may need to do from bash:
$ some.bat <Windows path>
If this is there are a lot of solutions including cygpath, doubling
backslashes, quoting file names or just using forward slashes (which
Windows often works just fine with).
> and simply because bash command lines are interpreted differently than
> Windows CMD command line (for example, when it comes to quoting or
> parameter substitution).
We all know that bash interprets things the bash way. And no doubt cmd
interprets things it's way. I do not see cmd's interpretation as superior.
> The main problem, however, is: If you are going to deliver something,
> which is supposed to run under CMD.EXE, most customers won't accept it
> until you really have tested it under CMD.EXE, and for good reason.
That's where:
$ cmd /c somescript.bat
comes in. That does run it under cmd.exe doesn't it? Why do you need
anything more? Why do you need to enter multiple commands and have
command history? Why not simply:
$ cmd /c testscript1.bat
<verify output>
$ # That worked!
$ cmd /c testscript2.bat
<verify output>
$ # That worked too...
$ # etc...
> In fact, even though I got running rxvt with cmd thanks to all the
> suggestions to my post, I will do some *final* tests still in plain,
> native Windoze Command-Windows, just for the safe side.
>
> Using a bash shell as a "main work horse" is great, but when you have
> the pleasure to create and test batch files, you will sooner or later be
> happy to also have a CMD shell available...
Just re-write your .bat scripts in Perl! ;-)
>
>> In both cases you generally to present Windows paths, of course;
>> cygpath can help with that.
> I use cygpath in several of my scripts and it's extremely useful, but
> dealing with the various path representations in interactive work is,
> for me at least, an annoyance...
It's really not that hard. Just s/\\/\/g! That often is enough. Or do a
cygpath -w <posix path> or whatever.
--
Andrew DeFaria <http://defaria.com>
My wife keeps complaining I never listen to her ...or something like that.
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
- Raw text -