Emmanuel Dreyfus | 4 Jul 12:58 2011
Picon

Re: bumping library major on release branch?

Emmanuel Dreyfus <manu <at> netbsd.org> wrote:

> Attached is a first attempt at it. I am not sure of how to use __RENAME()
> here. Any comment?

Hello

Nobody comented on this patch. Is it acceptable for a netbsd-5 pullup? 

--

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu <at> netbsd.org

Emmanuel Dreyfus | 4 Jul 14:08 2011
X-Face
Picon

Re: bumping library major on release branch?

On Mon, Jul 04, 2011 at 12:58:47PM +0200, Emmanuel Dreyfus wrote:
> Nobody comented on this patch. Is it acceptable for a netbsd-5 pullup? 

Here is it again, for a more convenient reference
http://ftp.espci.fr/shadow/manu/puffs_ea_lib.patch
--

-- 
Emmanuel Dreyfus
manu <at> netbsd.org

Manuel Bouyer | 4 Jul 15:29 2011

Re: bumping library major on release branch?

On Mon, Jul 04, 2011 at 12:08:28PM +0000, Emmanuel Dreyfus wrote:
> On Mon, Jul 04, 2011 at 12:58:47PM +0200, Emmanuel Dreyfus wrote:
> > Nobody comented on this patch. Is it acceptable for a netbsd-5 pullup? 
> 
> Here is it again, for a more convenient reference
> http://ftp.espci.fr/shadow/manu/puffs_ea_lib.patch

why _puffs_init52 and puffs_ops51 ?

shouldn't PUFFS_DEVEL_LIBVERSION also be changed ?

--

-- 
Manuel Bouyer <bouyer <at> antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--

Emmanuel Dreyfus | 4 Jul 15:49 2011
Picon

Re: bumping library major on release branch?

Manuel Bouyer <bouyer <at> antioche.eu.org> wrote:

> why _puffs_init52 and puffs_ops51 ?

_puffs_init52 because binaries from 5.1 are linked to used _puffs_init,
so the old _puffs_init should keep its old name, and I have to change
the newer one.

puffs_ops51 because when building a binary on 5.2, we will use puffs_ops
and we do not want to use the 5.1 structure. Therefore I have to change
the older structure name.

It also looks odd to me, I suspect I am missing something.

> shouldn't PUFFS_DEVEL_LIBVERSION also be changed ?

Not sure: when are we supposed to do that?

--

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu <at> netbsd.org

Manuel Bouyer | 4 Jul 15:46 2011

Re: bumping library major on release branch?

On Mon, Jul 04, 2011 at 03:49:40PM +0200, Emmanuel Dreyfus wrote:
> Manuel Bouyer <bouyer <at> antioche.eu.org> wrote:
> 
> > why _puffs_init52 and puffs_ops51 ?
> 
> _puffs_init52 because binaries from 5.1 are linked to used _puffs_init,
> so the old _puffs_init should keep its old name, and I have to change
> the newer one.
> 
> puffs_ops51 because when building a binary on 5.2, we will use puffs_ops
> and we do not want to use the 5.1 structure. Therefore I have to change
> the older structure name.

OK, it's because one refers to the new function, and the other to the old
structure.

>  
> It also looks odd to me, I suspect I am missing something.
> 
> > shouldn't PUFFS_DEVEL_LIBVERSION also be changed ?
> 
> Not sure: when are we supposed to do that?

Dunno, but the author can probably tell :)

--

-- 
Manuel Bouyer <bouyer <at> antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--

(Continue reading)

Emmanuel Dreyfus | 4 Jul 16:03 2011
Picon

Re: bumping library major on release branch?

Manuel Bouyer <bouyer <at> antioche.eu.org> wrote:

> OK, it's because one refers to the new function, and the other to the old
> structure.

It there a prefered way to do it?

> > > shouldn't PUFFS_DEVEL_LIBVERSION also be changed ?
> > Not sure: when are we supposed to do that?
> Dunno, but the author can probably tell :)

Um, PUFFS_DEVEL_LIBVERSION was dropped in -current. That suggests it is
not that important :-)

--

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu <at> netbsd.org

Manuel Bouyer | 4 Jul 15:58 2011

Re: bumping library major on release branch?

On Mon, Jul 04, 2011 at 04:03:26PM +0200, Emmanuel Dreyfus wrote:
> Manuel Bouyer <bouyer <at> antioche.eu.org> wrote:
> 
> > OK, it's because one refers to the new function, and the other to the old
> > structure.
> 
> It there a prefered way to do it?

I depends on how it's used in the API I guess.

> 
> > > > shouldn't PUFFS_DEVEL_LIBVERSION also be changed ?
> > > Not sure: when are we supposed to do that?
> > Dunno, but the author can probably tell :)
> 
> Um, PUFFS_DEVEL_LIBVERSION was dropped in -current. That suggests it is
> not that important :-)

OK, no need to touch it then.

--

-- 
Manuel Bouyer <bouyer <at> antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--

Emmanuel Dreyfus | 4 Jul 17:28 2011
Picon

Re: bumping library major on release branch?

Manuel Bouyer <bouyer <at> antioche.eu.org> wrote:

> OK, no need to touch it then.

I submitted a pullup with a combined patch with all extended attribute
stuff. It is huge, but the different changes sets depend on each other.
And at least the combined patch matches the code I have been testing.

--

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu <at> netbsd.org

Abhinav Upadhyay | 8 Jul 19:24 2011
Picon

[GSoC] [Apropos Replacement] Midterm Project Status Update

Hello NetBSD!!

I am writing to report about the status of my project as we are
approaching the midterm evaluation dates (11th July).

As always I have put up a detailed blog post about the changes and
improvements I did as I can format the content better there making it
easier to read. Here is the post:
http://abhinav-upadhyay.blogspot.com/2011/07/netbsd-gsoc-midterm-project-update.html

Following are the major changes I did:

1. Showing the section number along with search results: Last time
around, there was no section number with the search results, which
made the results slightly incomplete. Thanks to Kristaps about
suggesting the right method to extract the meta data out of the man
pages, this is fixed.

2. Improved the Ranking Algorithm By Including the Inverse Document
Frequency (IDF): Earlier, the ranking algorithm was only based on the
term frequency (the number of times a term appears in a document). But
I updated it to also include the IDF factor.
Inverse Document Frequency for a term is defined as the number of
documents in which the ter appears. Combining both tf and idf results
in improvement in the quality of results, as it now considers both how
often a term appears in a document, and also how often the term
appears in the whole corpus.

 IDF is calculated as: idf = log(n / nt)
Where n = Total number of documents in the corpus
(Continue reading)

Dustin Marquess | 9 Jul 02:03 2011
Picon

mksh

I see back in Dec head brought up the idea of replacing /bin/ksh with mksh:

http://mail-index.netbsd.org/tech-userlevel/2010/12/30/msg004328.html

Skimming through the discussion it seems like there was a consensus
that replacing /bin/sh would be a bad idea, but I don't see a decision
about /bin/ksh.  Is this still be decided, or is it basically dead?

I'm just curious, as /bin/ksh doesn't seem to be maintained that much
anymore, and especially compared to R40 of mksh, it seems to be
getting a little stale.

-Dustin


Gmane