[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Plink2 (Was: Bug#771154: plink: New major upstream version of plink)



Dear all,

On 09-01-15 09:18, Andreas Tille wrote:
> Hi Dylan,
> 
> the reason why I did not answered is that I'm finally no plink user and
> thus feel incompetent for a sensible hint here.

I am a plink user, so I'll give my two cents below.

>   In any case I have put
> the item "name space pollution" on my agenda for the upcoming Debian Med
> sprint[2].  (BTW, are there any more people interested in joining the
> sprint?  This is a really good oportunity to learn to know each other
> and strengthen cooperation.)

I'd love to come, but unfortunately I have other obligations during
those days.

> 
> On Tue, Jan 06, 2015 at 07:47:50AM +0100, Dylan wrote:
>> Hi Andreas,
>> I'm not very happy with the rename of the binary to "plink2" for this
>> version. I speak with upstream about this [1] and he suggest me to not
>> rename plink 1.9 to "plink2" because the real plink 2 will break more
>> things and will be distributed under the "plink2" name.

I agree.

>>
>> So the options are:
>> 1. Stay with the rename to "plink2" for the version 1.9. But with possible
>> breaks when the real plink2 will be released.
> 
> This could be carefully documented to warn users.

Given that plink 1.9 is actually still in beta phase (and as far as I
know is actually a beta version of plink 2), I wouldn't go for this option.

>  
>> 2. As suggest by upstream, stay with "p-link" for the version 1.9 with a
>> Breaks or Conflicts field in d/control for the plink 1.07. Not possible to
>> use both versions (1.07 and 1.9) at the same time, so not ideal to
>> replicate results obtained from plink 1.07.
> 
> I also think that this is not a good solution.  If I would use a new
> version of a scientifix program I would throw my data on both versions
> and compare the results.  We would force users to create virtual
> machines or something like this to follow this approach.  Does not sound
> very comfortable to me.

I agree.

> 
>> 3. Create a specific package ("plink1.9") for plink 1.9 to permit to have
>> the three versions in cohabitation during the transition. Which could
>> permit to replicate results obtained with all versions.
> 
> I'd probably choose this option.

Me too. Ideally with the option to use dpkg-reconfigure to allow the
sysadmin to set symlink that points p-link (or ideally plink) to one of
the three versions (like you can do with e.g. gcc versions).

Regarding the p-link vs plink vs plink1 vs plink19 vs plink2 naming of
the executable: I understand that the plink name is taken by the Putty
package, but I was wondering if the Debian guidelines allow to use a
mechanism like dpkg-reconfigure to allow the user to choose which of
these applications is actually pointed to when the user types plink.

My main argument for using plink (in fact as a sysadmin I always
manually create a plink symlink in /usr/local/bin that points to either
/usr/bin/p-link or the /usr/lib/plink/plink binary directly) is that all
tutorials and courses in genetic epidemiology that I know (including my
own) use the plink name. It is (obviously) what is used in the plink
documentation and since many people (in my experience) that are using
plink are not very experienced with the Linux command line I'd like to
keep things as simple and straight forward as possible.
Alternatively, would it be possible for the installer to ask a question
about whether to set this symlink or not (unless the variable
DEBIAN_FRONTEND is set to non-interactive), similar to e.g. the password
question when installing MySQL. Maybe such a question should only be
asked if the putty package is not installed.

>  
>> 4. Others solutions?
> 
> Hoping for more input here as well.
> 
> Kind regards and thanks for your work on this

Indeed, thanks for packaging this Dylan!


Best regards,

Lennart.

> 
>     Andreas.
> 
>  
>> [1] https://github.com/chrchang/plink-ng/issues/12
> 
> [2] https://wiki.debian.org/Sprints/2015/DebianMed2015#Agenda
> 

-- 
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
L.C. Karssen
Utrecht
The Netherlands

lennart@karssen.org
http://blog.karssen.org
GPG key ID: A88F554A
-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-

Attachment: signature.asc
Description: OpenPGP digital signature


Reply to: