More: [1.7] packaging problem? Both /usr/bin/ and /usr/lib/ are non-empty

Corinna Vinschen
Tue May 12 14:06:00 GMT 2009

On May 12 09:51, Charles Wilson wrote:
> Corinna Vinschen wrote:
> > this could only happen if /etc/fstab hasn't been created right at 
> > the start of the postinstall.
> Apparently that's exactly what has happened...
> > Did you examine setup.log.full at the point where the postinstall
> > starts?  What's the package order?  Does
> > C:\cygwin-1.7\bin\bash.exe -c /etc/postinstall/
> > run first?  Any error messages?
> ...
> Installing file cygfile:///etc/postinstall/
> 2009/05/11 14:55:57 Changing gid back to original
> Visited: 441 nodes out of 1454 while creating dependency order.
> Dependency order of packages: base-cygwin base-passwd cygwin libiconv2
> libintl8 libintl3 libgmp3 gawk tzcode coreutils texinfo _update-info-dir
> [... removed ...]

Looks good.

> 2009/05/11 14:55:58 running: C:\cygwin-1.7\bin\bash.exe -c
> /etc/postinstall/
> Huh?  No /etc/fstab file in \??\C:\cygwin-1.7\etc\fstab.d\cwilso11? 
> Using default root and cygdrive prefix...

See this message?  That means, at the time
was called the Cygwin DLL didn't find an fstab file.

> /cygdrive/h/.bashrc


> 2009/05/11 14:56:07 running: C:\cygwin-1.7\bin\bash.exe -c
> /etc/postinstall/

No such "Huh?" message anymore.  So at this point the /etc/fstab file
already exists.

> /cygdrive/h/.bashrc

But I don't get what this .bashrc pathname is doing here every time.

> I think I agree with Eric. I know what we're talking about, and even I
> was momentarily confused by the use of "read only" even in these
> discussions.  It's fine if you want to (re)use the currently
> non-functional MOUNT_RO option internally, as we (cygwin) can't control
> whether a mount point is read-only at all anyway. That's up to Windows. 
> But...the documentation and user error messages should *not* say
> anything about "read only" -- in those cases, I think we should use a
> different term.  Perhaps Eric's suggestion of "fixed", or maybe some
> other synonym that doesn't imply 'correcting a defect', like:

Ok, that makes two confused.  I assume I'm not confused and don't
understand the confusion for the simple reason that I created the text.
I'll change that to use another term.

> permanent, steady
> Synonyms:       anchored, attached, established, fast, firm, hitched,
> hooked, immobile, immotile, immovable, located, locked, made fast,
> nailed, quiet, rigid, rooted, secure, set, settled, situated, solid,
> stable, steadfast, stiff, still, tenacious, tight

Thanks for the selection.

> But, those objections are all stylistic. Functionally, I think your
> patch addresses the known issues reasonably.  (Although, for some reason
> my installation didn't create the /etc/fstab.  Your patch will probably
> paper over whatever that underlying reason is; that is, I believe that
> with your change, my installation would have worked properly...but *why*
> did the current system fail to create /etc/fstab? Dunno...)

But the above snippet from setup.log.full suggests differently.  The
/etc/fstab file must have been created, otherwise you would have had
"Huh?" messages all over the place.

Yes, I'm perplexed.


Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          cygwin AT cygwin DOT com
Red Hat

More information about the Cygwin-developers mailing list