Re: Why libquadmath0-4.7.2

2013-03-11 Thread Dave Korn
On 05/03/2013 16:22, Christopher Faylor wrote:
 On Tue, Mar 05, 2013 at 05:11:34PM +0100, Angelo Graziosi wrote:
 I have GCC-4.5.3 installed and I am not going to install the test 
 version 4.7.2-1, but setup.exe *wants* to install libquadmath0-4.7.2-1 
 even if I have selected Curr packages (which should exclude any 
 reference to 4.7.2 version...).

 Have we sure it isn't a packaging bug?
 
 The crude dependency handling in setup.exe isn't up to the task of
 allowing per-version dependencies.  So this probably came along because
 it's required for a newer version of gcc.

  When last we spoke (on the -apps list), it was suggested that the
dependencies should remain correct for the curr: version, and people
installing the test: version should manually install the required dependencies.

  If it would help, I could upload a 4.7.2-2 later tonight which restores java
and has the original 4.5.3-3 dependencies?

  (Many thanks to JonY for stepping up to the plate while I've been away.)

cheers,
  DaveK


--
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



Re: Why libquadmath0-4.7.2

2013-03-11 Thread JonY
On 3/12/2013 02:03, Dave Korn wrote:
 On 05/03/2013 16:22, Christopher Faylor wrote:
 On Tue, Mar 05, 2013 at 05:11:34PM +0100, Angelo Graziosi wrote:
 I have GCC-4.5.3 installed and I am not going to install the test 
 version 4.7.2-1, but setup.exe *wants* to install libquadmath0-4.7.2-1 
 even if I have selected Curr packages (which should exclude any 
 reference to 4.7.2 version...).

 Have we sure it isn't a packaging bug?

 The crude dependency handling in setup.exe isn't up to the task of
 allowing per-version dependencies.  So this probably came along because
 it's required for a newer version of gcc.
 
   When last we spoke (on the -apps list), it was suggested that the
 dependencies should remain correct for the curr: version, and people
 installing the test: version should manually install the required 
 dependencies.
 
   If it would help, I could upload a 4.7.2-2 later tonight which restores java
 and has the original 4.5.3-3 dependencies?

You'll be retaking maintainership again? I don't mind that, but please
push as much of the patches as possible upstream, it was kind of hard to
wrap my head around the patches to manually apply them.

Thanks.




signature.asc
Description: OpenPGP digital signature


Re: Why libquadmath0-4.7.2

2013-03-11 Thread Christopher Faylor
On Mon, Mar 11, 2013 at 06:03:08PM +, Dave Korn wrote:
On 05/03/2013 16:22, Christopher Faylor wrote:
 On Tue, Mar 05, 2013 at 05:11:34PM +0100, Angelo Graziosi wrote:
 I have GCC-4.5.3 installed and I am not going to install the test 
 version 4.7.2-1, but setup.exe *wants* to install libquadmath0-4.7.2-1 
 even if I have selected Curr packages (which should exclude any 
 reference to 4.7.2 version...).

 Have we sure it isn't a packaging bug?
 
The crude dependency handling in setup.exe isn't up to the task of
allowing per-version dependencies.  So this probably came along because
it's required for a newer version of gcc.

When last we spoke (on the -apps list), it was suggested that the
dependencies should remain correct for the curr: version, and people
installing the test: version should manually install the required
dependencies.

Dave, I've pinged you on the -apps list and you haven't responded.

cgf

--
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



Re: Why libquadmath0-4.7.2

2013-03-11 Thread Dave Korn
On 11/03/2013 23:21, Christopher Faylor wrote:
 On Mon, Mar 11, 2013 at 06:03:08PM +, Dave Korn wrote:
 On 05/03/2013 16:22, Christopher Faylor wrote:
 On Tue, Mar 05, 2013 at 05:11:34PM +0100, Angelo Graziosi wrote:
 I have GCC-4.5.3 installed and I am not going to install the test 
 version 4.7.2-1, but setup.exe *wants* to install libquadmath0-4.7.2-1 
 even if I have selected Curr packages (which should exclude any 
 reference to 4.7.2 version...).

 Have we sure it isn't a packaging bug?
 The crude dependency handling in setup.exe isn't up to the task of
 allowing per-version dependencies.  So this probably came along because
 it's required for a newer version of gcc.
 When last we spoke (on the -apps list), it was suggested that the
 dependencies should remain correct for the curr: version, and people
 installing the test: version should manually install the required
 dependencies.
 
 Dave, I've pinged you on the -apps list and you haven't responded.

  How recently?  I've got a backlog of 2400 unread mails there (and 8682 mails
here).  Yes, I've been away, but I can now commit to being back on the case;
the project I've been working on for the past god-knows-how-long is done.
I've already got back to work on upstream GCC and would like to get back to
work on Cygwin.

cheers,
  DaveK

--
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



Re: Why libquadmath0-4.7.2

2013-03-11 Thread Dave Korn
On 11/03/2013 22:12, JonY wrote:
 On 3/12/2013 02:03, Dave Korn wrote:
 On 05/03/2013 16:22, Christopher Faylor wrote:
 On Tue, Mar 05, 2013 at 05:11:34PM +0100, Angelo Graziosi wrote:
 I have GCC-4.5.3 installed and I am not going to install the test 
 version 4.7.2-1, but setup.exe *wants* to install libquadmath0-4.7.2-1 
 even if I have selected Curr packages (which should exclude any 
 reference to 4.7.2 version...).

 Have we sure it isn't a packaging bug?
 The crude dependency handling in setup.exe isn't up to the task of
 allowing per-version dependencies.  So this probably came along because
 it's required for a newer version of gcc.
   When last we spoke (on the -apps list), it was suggested that the
 dependencies should remain correct for the curr: version, and people
 installing the test: version should manually install the required 
 dependencies.

   If it would help, I could upload a 4.7.2-2 later tonight which restores 
 java
 and has the original 4.5.3-3 dependencies?
 
 You'll be retaking maintainership again? I don't mind that, but please
 push as much of the patches as possible upstream, it was kind of hard to
 wrap my head around the patches to manually apply them.

  I would like to resume maintaining it, subject to any discussions necessary.
 Being an upstream maintainer makes it easier for me to push patches.

cheers,
  DaveK



--
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



Why libquadmath0-4.7.2

2013-03-05 Thread Angelo Graziosi
I have GCC-4.5.3 installed and I am not going to install the test 
version 4.7.2-1, but setup.exe *wants* to install libquadmath0-4.7.2-1 
even if I have selected Curr packages (which should exclude any 
reference to 4.7.2 version...).


Have we sure it isn't a packaging bug?

 Angelo

--
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



Re: Why libquadmath0-4.7.2

2013-03-05 Thread Christopher Faylor
On Tue, Mar 05, 2013 at 05:11:34PM +0100, Angelo Graziosi wrote:
I have GCC-4.5.3 installed and I am not going to install the test 
version 4.7.2-1, but setup.exe *wants* to install libquadmath0-4.7.2-1 
even if I have selected Curr packages (which should exclude any 
reference to 4.7.2 version...).

Have we sure it isn't a packaging bug?

The crude dependency handling in setup.exe isn't up to the task of
allowing per-version dependencies.  So this probably came along because
it's required for a newer version of gcc.

--
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