diff mbox

[2/3] whois: use github repo as home page

Message ID fa924d0754ef385b1706721751e5f6f00e60c199.1502649259.git.baruch@tkos.co.il
State Changes Requested
Headers show

Commit Message

Baruch Siach Aug. 13, 2017, 6:34 p.m. UTC
The github repo is more informative than the list Marco's software on
the previous link.

Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
---
 package/whois/Config.in | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Thomas Petazzoni Aug. 14, 2017, 8:03 p.m. UTC | #1
Hello,

On Sun, 13 Aug 2017 21:34:18 +0300, Baruch Siach wrote:
> The github repo is more informative than the list Marco's software on
> the previous link.
> 
> Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
> Signed-off-by: Baruch Siach <baruch@tkos.co.il>

What about switching the upstream to be Github, instead of fetching
from Debian? The Github repo has all the releases available:
https://github.com/rfc1036/whois/releases.

Just make sure that the 5.2.14 tarball we were getting from Debian is
similar in contents to what we get from the Github repository for the
same version.

Thanks,

Thomas
Baruch Siach Aug. 15, 2017, 4:14 a.m. UTC | #2
Hi Thomas,

On Mon, Aug 14, 2017 at 10:03:54PM +0200, Thomas Petazzoni wrote:
> On Sun, 13 Aug 2017 21:34:18 +0300, Baruch Siach wrote:
> > The github repo is more informative than the list Marco's software on
> > the previous link.
> > 
> > Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
> > Signed-off-by: Baruch Siach <baruch@tkos.co.il>
> 
> What about switching the upstream to be Github, instead of fetching
> from Debian? The Github repo has all the releases available:
> https://github.com/rfc1036/whois/releases.
> 
> Just make sure that the 5.2.14 tarball we were getting from Debian is
> similar in contents to what we get from the Github repository for the
> same version.

But we generally prefer upstream provided tarballs, aren't we? Moreover, in 
this case the Debian repo provides a smaller .tar.xz, and an upstream computed 
SHA256 in the .dsc file. You don't get that from github.

baruch
Arnout Vandecappelle Aug. 15, 2017, 10:10 a.m. UTC | #3
On 15-08-17 06:14, Baruch Siach wrote:
> Hi Thomas,
> 
> On Mon, Aug 14, 2017 at 10:03:54PM +0200, Thomas Petazzoni wrote:
>> On Sun, 13 Aug 2017 21:34:18 +0300, Baruch Siach wrote:
>>> The github repo is more informative than the list Marco's software on
>>> the previous link.
>>>
>>> Cc: Gustavo Zacarias <gustavo@zacarias.com.ar>
>>> Signed-off-by: Baruch Siach <baruch@tkos.co.il>
>> What about switching the upstream to be Github, instead of fetching
>> from Debian? The Github repo has all the releases available:
>> https://github.com/rfc1036/whois/releases.
>>
>> Just make sure that the 5.2.14 tarball we were getting from Debian is
>> similar in contents to what we get from the Github repository for the
>> same version.
> But we generally prefer upstream provided tarballs, aren't we? Moreover, in 
> this case the Debian repo provides a smaller .tar.xz, and an upstream computed 
> SHA256 in the .dsc file. You don't get that from github.

 On Sunday, I was going to ask the same question, and came to the same
conclusion about the tar.xz and the sha256. Unfortunately, I don't remember why
I didn't just apply it then and there...

 Regards,
 Arnout
Thomas Petazzoni Aug. 15, 2017, 10:56 a.m. UTC | #4
Hello,

On Tue, 15 Aug 2017 12:10:17 +0200, Arnout Vandecappelle wrote:

> > But we generally prefer upstream provided tarballs, aren't we? Moreover, in 
> > this case the Debian repo provides a smaller .tar.xz, and an upstream computed 
> > SHA256 in the .dsc file. You don't get that from github.  
> 
>  On Sunday, I was going to ask the same question, and came to the same
> conclusion about the tar.xz and the sha256. Unfortunately, I don't remember why
> I didn't just apply it then and there...

I definitely don't agree here. We should use upstream when it exists,
even if it provides things that are less "convenient" than Debian.

Otherwise, we're going to convert a *lot* of packages to use tarballs
from Debian rather than from upstream.

What will happen when we'll want to bump to a new upstream version that
hasn't been packaged by Debian ?

Sorry, but I definitely disagree here. Please use the real upstream.

Best regards,

Thomas
Baruch Siach Aug. 15, 2017, 11:20 a.m. UTC | #5
Hi Thomas,

On Tue, Aug 15, 2017 at 12:56:21PM +0200, Thomas Petazzoni wrote:
> On Tue, 15 Aug 2017 12:10:17 +0200, Arnout Vandecappelle wrote:
> 
> > > But we generally prefer upstream provided tarballs, aren't we? Moreover, in 
> > > this case the Debian repo provides a smaller .tar.xz, and an upstream computed 
> > > SHA256 in the .dsc file. You don't get that from github.  
> > 
> >  On Sunday, I was going to ask the same question, and came to the same
> > conclusion about the tar.xz and the sha256. Unfortunately, I don't remember why
> > I didn't just apply it then and there...
> 
> I definitely don't agree here. We should use upstream when it exists,
> even if it provides things that are less "convenient" than Debian.
> 
> Otherwise, we're going to convert a *lot* of packages to use tarballs
> from Debian rather than from upstream.
> 
> What will happen when we'll want to bump to a new upstream version that
> hasn't been packaged by Debian ?

As things stand now this is quite unlikely. The whois package Debian 
maintainer and the upstream developer are the same person, Marco d'Itri. The 
only upstream in-tree changelog file is at debian/changelog.

> Sorry, but I definitely disagree here. Please use the real upstream.

The real whois upstream README file says this:

  The canonical distribution point for releases of the program is
  http://ftp.debian.org/debian/pool/main/w/whois/ .

baruch
Thomas Petazzoni Aug. 15, 2017, 12:07 p.m. UTC | #6
Hello,

On Tue, 15 Aug 2017 14:20:18 +0300, Baruch Siach wrote:

> On Tue, Aug 15, 2017 at 12:56:21PM +0200, Thomas Petazzoni wrote:
> > On Tue, 15 Aug 2017 12:10:17 +0200, Arnout Vandecappelle wrote:
> >   
> > > > But we generally prefer upstream provided tarballs, aren't we? Moreover, in 
> > > > this case the Debian repo provides a smaller .tar.xz, and an upstream computed 
> > > > SHA256 in the .dsc file. You don't get that from github.    
> > > 
> > >  On Sunday, I was going to ask the same question, and came to the same
> > > conclusion about the tar.xz and the sha256. Unfortunately, I don't remember why
> > > I didn't just apply it then and there...  
> > 
> > I definitely don't agree here. We should use upstream when it exists,
> > even if it provides things that are less "convenient" than Debian.
> > 
> > Otherwise, we're going to convert a *lot* of packages to use tarballs
> > from Debian rather than from upstream.
> > 
> > What will happen when we'll want to bump to a new upstream version that
> > hasn't been packaged by Debian ?  
> 
> As things stand now this is quite unlikely. The whois package Debian 
> maintainer and the upstream developer are the same person, Marco d'Itri. The 
> only upstream in-tree changelog file is at debian/changelog.
> 
> > Sorry, but I definitely disagree here. Please use the real upstream.  
> 
> The real whois upstream README file says this:
> 
>   The canonical distribution point for releases of the program is
>   http://ftp.debian.org/debian/pool/main/w/whois/ .

Fair enough. But I still find it completely weird to have the Github
page as the "project home page" indicated in the Config.in file, and
not use that as the upstream.

And Arnout has put PATCH 1/3 back to the New state, while I still
disagree with the approach being taken.

Best regards,

Thomas
Arnout Vandecappelle Aug. 15, 2017, 3:23 p.m. UTC | #7
On 15-08-17 14:07, Thomas Petazzoni wrote:
> Hello,
> 
> On Tue, 15 Aug 2017 14:20:18 +0300, Baruch Siach wrote:
> 
>> On Tue, Aug 15, 2017 at 12:56:21PM +0200, Thomas Petazzoni wrote:
[snip]
>> The real whois upstream README file says this:
>>
>>   The canonical distribution point for releases of the program is
>>   http://ftp.debian.org/debian/pool/main/w/whois/ .
> 
> Fair enough. But I still find it completely weird to have the Github
> page as the "project home page" indicated in the Config.in file, and
> not use that as the upstream.

 It's weird, but it is what it is. Don't we have any other package where the
github repo is the upstream URL, but the release tarball comes from somewhere else?


> And Arnout has put PATCH 1/3 back to the New state, while I still
> disagree with the approach being taken.

 Sorry, I hadn't seen your reply to 1/3, I had also seen your reply to 2/3 and
considered that Baruch and I had given sufficient feedback to warrant going back
to New. With your reply to 1/3, it indeed makes sense to set the entire series
to Changes Requested.

 However, I'm just going to apply patch 3/3 now, since there is no objection to
that one. So only 1/3 and 2/3 will be Changes Requested.
diff mbox

Patch

diff --git a/package/whois/Config.in b/package/whois/Config.in
index 9505d5832056..d2a030282b76 100644
--- a/package/whois/Config.in
+++ b/package/whois/Config.in
@@ -4,4 +4,4 @@  config BR2_PACKAGE_WHOIS
 	help
 	  Improved whois client.
 
-	  http://www.linux.it/~md/software/
+	  https://github.com/rfc1036/whois