Re: Cdbs Features

2019-05-22 Thread Marco d'Itri
On May 22, Wouter Verhelst  wrote:

> I think at this point we can recommend dh, and require debhelper (i.e.,
> the individual dh_* tools could be required to be part of the build
> system, but how they are called can be left to a maintainer's
> discretion, with the assumption that "you use dh or perhaps cdbs, except
> if you're trying to develop something better, but then you still use
> debhelper behind the scenes to do the low-level work"
This looks reasonable to me too.

(Now I am almost sold on dh, BTW.)

-- 
ciao,
Marco


signature.asc
Description: PGP signature


Re: Cdbs Features

2019-05-22 Thread Wouter Verhelst
On Wed, May 15, 2019 at 02:37:59PM +0900, Marc Dequènes (duck) wrote:
> On the more general topic, I believe there should be room for new tools to
> emerge and not-being-dh should never be a RC or even important bug.

My thoughts exactly.

I think at this point we can recommend dh, and require debhelper (i.e.,
the individual dh_* tools could be required to be part of the build
system, but how they are called can be left to a maintainer's
discretion, with the assumption that "you use dh or perhaps cdbs, except
if you're trying to develop something better, but then you still use
debhelper behind the scenes to do the low-level work"

I'll note that cdbs also uses the low-level debhelper commands behind
the scenes.

-- 
To the thief who stole my anti-depressants: I hope you're happy

  -- seen somewhere on the Internet on a photo of a billboard



Re: Cdbs Features

2019-05-14 Thread duck

Quack,

On 2019-05-14 11:24, Sean Whitton wrote:

Switching the entire Haskell ecosystem over to use dh would be a 
massive

amount of work, as the new dh_haskell would need a lot of testing etc.
So Haskell libraries and apps would probably have to be one of the
exceptions.


It took months to get somethings working for Python and Ruby and in the 
end years to polish all the things and migrate all packages, so it's 
clearly going to take quite some time.


I think one of the major feature of CDBS is the way stages are grouped 
in make rules and you can hook onto it or even override. Coupled with 
patsubst it makes some kind of loops you can use to iterate over groups 
of packages, which is very handy when you need to build for various 
versions of an interpreter. So unlike DH you can really add or override 
parts of these loops if your package requires it. But in the end all 
these rules are complex to read and maintain, and if they were very 
useful in the past they also pushed debhelper/dh into going forward with 
more bold features and I don't think CDBS adds much to the table 
nowadays.


I would also point out that the maintenance of CDBS has been on Jonas 
Smedegaard's shoulders for years now and it has been difficult having 
almost no backup (which I'm partly responsible of). I think I was mostly 
able to convince him we should work on a retirement plan but I've had no 
time to do anything about it. Bugs like #885407 and others are still 
around while we are close to release, thus I still think this is the 
best course of action.



On the more general topic, I believe there should be room for new tools 
to emerge and not-being-dh should never be a RC or even important bug. 
Nevertheless I think this tool has grown well and a strong 
recommendation would be fine. I believe as a project we could agree on 
major goals around deprecating tools that lost traction and proper 
maintenance, old versions of the tools, old patch formats and so on, and 
encouraging contribution around it. We could also agree on some practice 
to avoid like direct source patching (especially when not on a VCS). But 
this way you're still free to use the patching system that suits you or 
build your own and suggest it to the community; that's how many of our 
tools were built.


As for using NMU I think this is a bad idea. Even if we're supposed to 
care after a NMU, this is no long-term maintenance and a returning busy 
maintainer would not really appreciate this. I believe adoption, 
becoming co-maintainer or creating a team would be better if you want to 
make drastic changes on a package and your one-off patch lingers in the 
BTS.


\_o<

--
Marc Dequènes



Re: Cdbs Features

2019-05-14 Thread Tobias Frost
On Tue, May 14, 2019 at 05:35:21AM -0700, Mo Zhou wrote:
> Hi,
> 
> On 2019-05-14 07:59, IOhannes m zmölnig wrote:
> > i've migrated many of my packages from cdbs to dh, but there's one
> > feature which cdbs sports and which i miss strongly (at least: the last
> > time i checked) in dh (so much, that i haven't converted a couple of
> > packages): building of multiple "flavours".
> > that is: building the same code-base multiple times, with differing
> > configurations.
> 
> Could you please provide some examples about the case you mentioned?
> I need exactly the "multiple flavours" feature for at least 2 packages:

"check" is compiling 2 flavours with dh:
https://tracker.debian.org/media/packages/c/check/rules-0.12.0-0.1
> 



Re: Cdbs Features [and 1 more messages]

2019-05-14 Thread Ian Jackson
Mo Zhou writes ("Re: Cdbs Features"):
> On 2019-05-14 07:59, IOhannes m zmölnig wrote:
> > that is: building the same code-base multiple times, with differing
> > configurations.
> 
> Could you please provide some examples about the case you mentioned?
> I need exactly the "multiple flavours" feature for at least 2 packages:
> 
> https://salsa.debian.org/science-team/blis/blob/master/debian/rules
> https://salsa.debian.org/science-team/openblas/blob/lumin/debian/rules
...
> I'm quite interested in taking a look at how cdbs deal with such case.

IOhannes m zmölnig (Debian/GNU) writes ("Re: Cdbs Features"):
> On 14.05.19 14:35, Mo Zhou wrote:
> > I'm quite interested in taking a look at how cdbs deal with such case.
> 
> https://salsa.debian.org/multimedia-team/snd/blob/master/debian/rules
> https://salsa.debian.org/multimedia-team/soundscaperenderer/blob/master/debian/rules

Wow.  This is like night and day.  My instinct is definitely that I
don't like cdbs but omg dh needs this feature.

I looked through the src:debhelper bugs and didn't see a request for
this.

Ian.

-- 
Ian JacksonThese opinions are my own.

If I emailed you from an address @fyvzl.net or @evade.org.uk, that is
a private address which bypasses my fierce spamfilter.



Re: Cdbs Features

2019-05-14 Thread Debian/GNU
On 14.05.19 14:35, Mo Zhou wrote:
> I'm quite interested in taking a look at how cdbs deal with such case.


https://salsa.debian.org/multimedia-team/snd/blob/master/debian/rules
https://salsa.debian.org/multimedia-team/soundscaperenderer/blob/master/debian/rules

gfmdsrt
IOhannes



Re: Cdbs Features

2019-05-14 Thread Mo Zhou
Hi,

On 2019-05-14 07:59, IOhannes m zmölnig wrote:
> i've migrated many of my packages from cdbs to dh, but there's one
> feature which cdbs sports and which i miss strongly (at least: the last
> time i checked) in dh (so much, that i haven't converted a couple of
> packages): building of multiple "flavours".
> that is: building the same code-base multiple times, with differing
> configurations.

Could you please provide some examples about the case you mentioned?
I need exactly the "multiple flavours" feature for at least 2 packages:

https://salsa.debian.org/science-team/blis/blob/master/debian/rules
https://salsa.debian.org/science-team/openblas/blob/lumin/debian/rules

On amd64 the two source packages will be compiled 6 times in order to
produce 6 different variants.

I'm quite interested in taking a look at how cdbs deal with such case.

Thanks.



Re: Cdbs Features

2019-05-14 Thread Debian/GNU



On 13.05.19 18:22, Sam Hartman wrote:
>> "Holger" == Holger Levsen  writes:
> Holger> - packages using cdbs. cdbs has features dh doesnt have and
> Holger> I dont think it's wrong to use cdbs. (
> 
> Just for my information, what are the big features cdbs has that dh does
> not?
> 

i've migrated many of my packages from cdbs to dh, but there's one
feature which cdbs sports and which i miss strongly (at least: the last
time i checked) in dh (so much, that i haven't converted a couple of
packages): building of multiple "flavours".
that is: building the same code-base multiple times, with differing
configurations.

while it is possible to do that with dh, it amounts to either a lot of
near-duplicate lines or quite some GNU make trickery, which i'd rather
not spread across multiple d/rules


fgamsdr
IOhannes



Re: Cdbs Features

2019-05-13 Thread Sean Whitton
Hello Sam,

On Mon 13 May 2019 at 12:22PM -04, Sam Hartman wrote:

>> "Holger" == Holger Levsen  writes:
> Holger> - packages using cdbs. cdbs has features dh doesnt have and
> Holger> I dont think it's wrong to use cdbs. (
>
> Just for my information, what are the big features cdbs has that dh does
> not?

I don't know cdbs well, so I don't know what big features it might have
that dh does not have, but I do know that the Haskell package ecosystem
relies on cdbs addon code.  I.e. in our d/rules we have

include /usr/share/cdbs/1/rules/debhelper.mk
include /usr/share/cdbs/1/class/hlibrary.mk

There have been attempts to rewrite what hlibrary.mk does as a
'dh_haskell' tool.  These attempts stalled.

Switching the entire Haskell ecosystem over to use dh would be a massive
amount of work, as the new dh_haskell would need a lot of testing etc.
So Haskell libraries and apps would probably have to be one of the
exceptions.

-- 
Sean Whitton


signature.asc
Description: PGP signature


Cdbs Features

2019-05-13 Thread Sam Hartman
> "Holger" == Holger Levsen  writes:
Holger> - packages using cdbs. cdbs has features dh doesnt have and
Holger> I dont think it's wrong to use cdbs. (

Just for my information, what are the big features cdbs has that dh does
not?