Re: ArrayExpressHTS ran into removed fastx-toolkit

[Steffen Möller]

On 28.07.21 06:30, Andreas Tille wrote:

Hi,

On Wed, Jul 28, 2021 at 01:13:38PM +0900, Charles Plessy wrote:

I had a quick lookKind regards

   Andreas.


  (git grep) at the source code of ArrayExpressHTS and
although it has an option to indicate the path to fasta_formatter, it
appears to never use it.

I patched it out. Many thanks for spotting that!

... and may be report this issue to ArrayExpressHTS upstream?



With that boost, I also had a look at patching out tophat but it seems
like that protocol is much depending on all the bits and pieces that
were nice and shiny when the package was introduced in 2010. We should
discuss with upstream what to do about the package. At the moment, the
only values I see are
 a) retrieving data from ArrayExpress for RNA-seq data (judge
yourselves on
https://www.bioconductor.org/packages/release/bioc/vignettes/ArrayExpressHTS/inst/doc/ArrayExpressHTS.R)
which was my motivation to package it
 b) increasing %age of what we cover from Bioconductor
 c) increasing %age of what we cover from CRAN

Best,
Steffen

Re: ArrayExpressHTS ran into removed fastx-toolkit

[Andreas Tille]

Hi,

On Wed, Jul 28, 2021 at 01:13:38PM +0900, Charles Plessy wrote:
> 
> I had a quick look (git grep) at the source code of ArrayExpressHTS and
> although it has an option to indicate the path to fasta_formatter, it
> appears to never use it.

... and may be report this issue to ArrayExpressHTS upstream?
 
Kind regards

  Andreas. 

-- 
http://fam-tille.de

Re: ArrayExpressHTS ran into removed fastx-toolkit

[Charles Plessy]

Le Tue, Jul 27, 2021 at 12:02:00PM +0200, Steffen Möller a écrit :
> 
> fastx-toolkit https://tracker.debian.org/pkg/fastx-toolkit ran into a
> bug in one of its dependencies and was removed for that reason.
> ArrayExpressHTS still depends on fasta_formatter from that library.

Hi Steffen,

I had a quick look (git grep) at the source code of ArrayExpressHTS and
although it has an option to indicate the path to fasta_formatter, it
appears to never use it.

As a control, you can git grep other external tools (samtools, mmseq,
...) and you will see that, in contrary to the search with
fasta_formatter, it returns lines of code that do construct a shell
command using the path to the tool.

So if I am not mistaken, maybe we can just drop its dependancy ?

Have a nice day,

Charles

-- 
Charles Plessy Nagahama, Yomitan, Okinawa, Japan
Debian Med packaging team http://www.debian.org/devel/debian-med
Tooting from work,   https://mastodon.technology/@charles_plessy
Tooting from home, https://framapiaf.org/@charles_plessy

Re: ArrayExpressHTS ran into removed fastx-toolkit

[Pirate Praveen]

On 2021, ജൂലൈ 27 4:28:53 PM IST, "Steffen Möller"  
wrote:
>
>On 27.07.21 12:53, Nilesh Patra wrote:
>>
>>
>> On Tue, 27 Jul, 2021, 4:16 pm Nilesh Patra, > > wrote:
>>
>> and its reverse dependencies
>> like ArrayExpressHTS with it, then?
>>
>>
>> Not needed for reverse dependencies, if we push those to ofiicial
>> bullseye-backports.
>>
>>     
>> Erm, correction here. Ofcourse, we will need to push in reverse-deps
>> of fastx-toolkit to fasttrack.
>>
>> But I think in such a case we might end up pushing relatively more
>> useful (chain of) packages in there.
>> But everything depends on if fastx-toolkit is absolutely impossible to
>> have in testing.
>> We do have several softwares in the archive where the upstream is
>> dead. Unless fastx-toolkit is completely not suited for release, we
>> could think about maintaining it ourselves
>
>And in my mind, keeping such older software afloat is part of the
>service that Debian provides to the community. So, let's hope for a path
>back into testing.

If the bug is fixed, it can get back into testing. If upstream is dead, how can 
you support in fast track ? Packages in fast track are still supported by 
upstream, they just don't support it long enough to cover a stable release 
cycle. For example, gitlab provides security support for 3 months only for a 
release, after that you need to upgrade to next feature release.

>Best,
>Steffen
>

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.

Re: ArrayExpressHTS ran into removed fastx-toolkit

[Pirate Praveen]

On 2021, ജൂലൈ 27 4:16:21 PM IST, Nilesh Patra  wrote:
>[Bringing Praveen into the loop who maintains fasttrack]
>
>Hi Steffen,
>
>On Tue, 27 Jul, 2021, 3:32 pm Steffen Möller, 
>wrote:
>
>> Hello,
>>
>> fastx-toolkit https://tracker.debian.org/pkg/fastx-toolkit ran into a
>> bug in one of its dependencies and was removed for that reason.
>> ArrayExpressHTS still depends on fasta_formatter from that library.
>>
>> In the meantime, Pierre has fixed that bug on that dependency of
>> fastx-toolkit, so I think this could go back in, right?
>>
>
>Sure, it could be reintroduced.
>
>If we are concerned about taking responsibility for fastx-toolkit since
>> upstream officially declared it as unsupported, then maybe this would be
>> something for https://fasttrack.debian.net/
>
>
>No, that's likely not the case for fasttrack. fasttrack is for packages
>that will never make it to testing. If it's the case for fastx-toolkit,
>then probably we go the fasttrack way
>
>and its reverse dependencies
>> like ArrayExpressHTS with it, then?
>
>
>Not needed for reverse dependencies, if we push those to ofiicial
>bullseye-backports
>
>But since complicates everything, I
>> would like not go that route.
>>
>
>Right. @Praveen, do you think it qualifies for being in fasttrack?

You are correct about it, fast track is an option if it can never be in stable, 
testing or back ports.
>Nilesh
>
>>

-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.

Re: ArrayExpressHTS ran into removed fastx-toolkit

[Steffen Möller]

On 27.07.21 12:53, Nilesh Patra wrote:



On Tue, 27 Jul, 2021, 4:16 pm Nilesh Patra, mailto:nil...@debian.org>> wrote:

and its reverse dependencies
like ArrayExpressHTS with it, then?


Not needed for reverse dependencies, if we push those to ofiicial
bullseye-backports.

    
Erm, correction here. Ofcourse, we will need to push in reverse-deps
of fastx-toolkit to fasttrack.

But I think in such a case we might end up pushing relatively more
useful (chain of) packages in there.
But everything depends on if fastx-toolkit is absolutely impossible to
have in testing.
We do have several softwares in the archive where the upstream is
dead. Unless fastx-toolkit is completely not suited for release, we
could think about maintaining it ourselves


And in my mind, keeping such older software afloat is part of the
service that Debian provides to the community. So, let's hope for a path
back into testing.

Best,
Steffen

Re: ArrayExpressHTS ran into removed fastx-toolkit

[Nilesh Patra]

On Tue, 27 Jul, 2021, 4:16 pm Nilesh Patra,  wrote:

> and its reverse dependencies
>> like ArrayExpressHTS with it, then?
>
>
> Not needed for reverse dependencies, if we push those to ofiicial
> bullseye-backports.
>

Erm, correction here. Ofcourse, we will need to push in reverse-deps of
fastx-toolkit to fasttrack.

But I think in such a case we might end up pushing relatively more useful
(chain of) packages in there.

> But everything depends on if fastx-toolkit is absolutely impossible to
have in testing.
We do have several softwares in the archive where the upstream is dead.
Unless fastx-toolkit is completely not suited for release, we could think
about maintaining it ourselves


>>

Re: ArrayExpressHTS ran into removed fastx-toolkit

[Nilesh Patra]

[Bringing Praveen into the loop who maintains fasttrack]

Hi Steffen,

On Tue, 27 Jul, 2021, 3:32 pm Steffen Möller, 
wrote:

> Hello,
>
> fastx-toolkit https://tracker.debian.org/pkg/fastx-toolkit ran into a
> bug in one of its dependencies and was removed for that reason.
> ArrayExpressHTS still depends on fasta_formatter from that library.
>
> In the meantime, Pierre has fixed that bug on that dependency of
> fastx-toolkit, so I think this could go back in, right?
>

Sure, it could be reintroduced.

If we are concerned about taking responsibility for fastx-toolkit since
> upstream officially declared it as unsupported, then maybe this would be
> something for https://fasttrack.debian.net/


No, that's likely not the case for fasttrack. fasttrack is for packages
that will never make it to testing. If it's the case for fastx-toolkit,
then probably we go the fasttrack way

and its reverse dependencies
> like ArrayExpressHTS with it, then?


Not needed for reverse dependencies, if we push those to ofiicial
bullseye-backports

But since complicates everything, I
> would like not go that route.
>

Right. @Praveen, do you think it qualifies for being in fasttrack?

Nilesh

>

ArrayExpressHTS ran into removed fastx-toolkit

[Steffen Möller]

Hello,

fastx-toolkit https://tracker.debian.org/pkg/fastx-toolkit ran into a
bug in one of its dependencies and was removed for that reason.
ArrayExpressHTS still depends on fasta_formatter from that library.

In the meantime, Pierre has fixed that bug on that dependency of
fastx-toolkit, so I think this could go back in, right?

If we are concerned about taking responsibility for fastx-toolkit since
upstream officially declared it as unsupported, then maybe this would be
something for https://fasttrack.debian.net/ and its reverse dependencies
like ArrayExpressHTS with it, then? But since complicates everything, I
would like not go that route.

Steffen