cygwin-pkg-maint maintance

cygwin-pkg-maint maintance

[Marco Atzeri]

Hi Maintainers,
I was looking at the needed maintenance of cygwin-pkg-maint,
crossing the data of x86 setup.ini with cygwin-pkg-maint
and I found 2 main classes of mismatch plus a minor one

1) missing packages.
Package that are in x86 setup.ini but not reported in cygwin-pkg-maint
at any level.
Nor the package, nor the source package and neither the first
subdirectory of x86/release is reported in the file;
I estimated the maintainer from the mailing list to be:

PACKAGE                 SOURCE_PKG              MAINTAINER

cloog-isl               cloog-isl               Achim Gratz
db4.5                   db4.5                   Dr. Volker Zell
db4.8                   db4.8                   Dr. Volker Zell
db4.8-doc               db4.8                   Dr. Volker Zell
font-misc-ethiopic      font-misc-ethiopic      Jon Turney
glib1.2                 glib1.2                 Yaakov Selkowitz
gtk1.2                  gtk1.2                  OBSOLETE (Yaakov Selkowitz)
gtk1.2-engines          gtk1.2-engines          OBSOLETE (Yaakov Selkowitz)
html2ps                 html2ps                 Yaakov Selkowitz
isl                     isl                     Achim Gratz
libart_lgpl_2           libart_lgpl_2           Yaakov Selkowitz
libaudio-devel          nas                     Yaakov Selkowitz
libaudio2               nas                     Yaakov Selkowitz
libcloog-isl-devel      cloog-isl               Achim Gratz
libcloog-isl-doc        cloog-isl               Achim Gratz
libcloog-isl4           cloog-isl               Achim Gratz
libedit                 libedit                 Corinna Vinschen
libedit-devel           libedit                 Corinna Vinschen
libedit0                libedit                 Corinna Vinschen
libisl-devel            isl                     Achim Gratz
libisl-doc              isl                     Achim Gratz
libisl10                isl                     Achim Gratz
libzzip-devel           zziplib                 Yaakov Selkowitz
libzzip0.13             zziplib                 Yaakov Selkowitz
mesa-demos              mesa-demos              Yaakov Selkowitz
nas                     nas                     Yaakov Selkowitz
nc                      nc                      Corinna Vinschen
printproto              printproto              OBSOLETE (Yaakov Selkowitz)
pwget                   pwget                   Jari Aalto
xhtml2ps                html2ps                 Yaakov Selkowitz
xlsx2csv                xlsx2csv                Jari Aalto
zziplib                 zziplib                 Yaakov Selkowitz

2) packages that are in cygwin-pkg-maint but not in setup.ini (nor x86 
or x86_64 ones)

most are OBSOLETE but not clearly defined as such, like
db2, db3.1, .. of Volker
libqhull, liblapack of mine


3) typos
	libgck1-ddoc -> libgck1-doc
	libzziplib-devel -> libzzip-devel
	

Question before I start to modify cygwin-pkg-maint:

Does still make sense to report all binary packages in
the cygwin-pkg-maint files ?
Should we cluster them at source package or directory level as accepted
by the upload procedure ?

Specially for the dll libraries the name are continuously bumped, and
we will be always behind. At first glance we are missing ~ 200
packages; so roughly 8% of all the binary packages (~ 2700).
The active source packages are ~ 1400 and they change much
less frequently (ITP, ITA, orphanage and obsolescence)
so a better accuracy is possible.

Regards
Marco

Re: cygwin-pkg-maint maintance

[Yaakov Selkowitz]

On 2014-08-10 15:33, Marco Atzeri wrote:
> I was looking at the needed maintenance of cygwin-pkg-maint,
> crossing the data of x86 setup.ini with cygwin-pkg-maint
> and I found 2 main classes of mismatch plus a minor one

Thanks for taking the time to look into this!

> 1) missing packages.
> Package that are in x86 setup.ini but not reported in cygwin-pkg-maint
> at any level.
> Nor the package, nor the source package and neither the first
> subdirectory of x86/release is reported in the file;
> I estimated the maintainer from the mailing list to be:
>
> PACKAGE                 SOURCE_PKG              MAINTAINER
>
> cloog-isl               cloog-isl               Achim Gratz
> db4.5                   db4.5                   Dr. Volker Zell
> db4.8                   db4.8                   Dr. Volker Zell
> db4.8-doc               db4.8                   Dr. Volker Zell
> font-misc-ethiopic      font-misc-ethiopic      Jon Turney
> glib1.2                 glib1.2                 Yaakov Selkowitz
> gtk1.2                  gtk1.2                  OBSOLETE (Yaakov Selkowitz)
> gtk1.2-engines          gtk1.2-engines          OBSOLETE (Yaakov Selkowitz)

The new upload mechanism doesn't seem to handle OBSOLETE well, so these 
two should be in my name without qualification in case I need to update 
them in the future.

> html2ps                 html2ps                 Yaakov Selkowitz
> isl                     isl                     Achim Gratz
> libart_lgpl_2           libart_lgpl_2           Yaakov Selkowitz
> libaudio-devel          nas                     Yaakov Selkowitz
> libaudio2               nas                     Yaakov Selkowitz
> libcloog-isl-devel      cloog-isl               Achim Gratz
> libcloog-isl-doc        cloog-isl               Achim Gratz
> libcloog-isl4           cloog-isl               Achim Gratz
> libedit                 libedit                 Corinna Vinschen
> libedit-devel           libedit                 Corinna Vinschen
> libedit0                libedit                 Corinna Vinschen
> libisl-devel            isl                     Achim Gratz
> libisl-doc              isl                     Achim Gratz
> libisl10                isl                     Achim Gratz
> libzzip-devel           zziplib                 Yaakov Selkowitz
> libzzip0.13             zziplib                 Yaakov Selkowitz
> mesa-demos              mesa-demos              Yaakov Selkowitz
> nas                     nas                     Yaakov Selkowitz
> nc                      nc                      Corinna Vinschen
> printproto              printproto              OBSOLETE (Yaakov Selkowitz)
> pwget                   pwget                   Jari Aalto
> xhtml2ps                html2ps                 Yaakov Selkowitz
> xlsx2csv                xlsx2csv                Jari Aalto
> zziplib                 zziplib                 Yaakov Selkowitz

The source packages should definitely be added; the subpackages depends 
on your last question.

> 2) packages that are in cygwin-pkg-maint but not in setup.ini (nor x86
> or x86_64 ones)
>
> most are OBSOLETE but not clearly defined as such, like
> db2, db3.1, .. of Volker
> libqhull, liblapack of mine

Do you have a complete list?

> 3) typos
>      libgck1-ddoc -> libgck1-doc
>      libzziplib-devel -> libzzip-devel

Yes, these are obvious fixes.

> Question before I start to modify cygwin-pkg-maint:
>
> Does still make sense to report all binary packages in
> the cygwin-pkg-maint files ?
> Should we cluster them at source package or directory level as accepted
> by the upload procedure ?
> Specially for the dll libraries the name are continuously bumped, and
> we will be always behind. At first glance we are missing ~ 200
> packages; so roughly 8% of all the binary packages (~ 2700).
> The active source packages are ~ 1400 and they change much
> less frequently (ITP, ITA, orphanage and obsolescence)
> so a better accuracy is possible.

I've actually been thinking along the same lines.  AFAICS, the new 
upload mechanism doesn't care about subpackage listings so long as the 
source package is listed (provided, of course, that the standard 
directory layout is used).

Therefore, the only "benefit" to having subpackages listed would be 
purely informative, but as you point out, 1) right now it's behind, 2) 
it's high maintenance for little gain, plus 3) a single person must 
anyway maintain an entire package with all its subpackages, and 4) nor 
does it actually indicate which subpackages map to which source packages.

So I am all in favour of making this a source-only list, in line with 
other distro package maintainer "databases".

Marco, if you're interested, please go ahead and add the missing 
*source* packages and fix those typos, but please hold off on removing 
anything until we hear from Corinna.

Thanks,

Yaakov

Re: cygwin-pkg-maint maintance

[Marco Atzeri]

On 11/08/2014 06:55, Yaakov Selkowitz wrote:
> On 2014-08-10 15:33, Marco Atzeri wrote:
>> I was looking at the needed maintenance of cygwin-pkg-maint,
>> crossing the data of x86 setup.ini with cygwin-pkg-maint
>> and I found 2 main classes of mismatch plus a minor one
>
> Thanks for taking the time to look into this!

It was a good excuse to refresh my python knowledge to cross the
two files data.


>> 1) missing packages.
>> Package that are in x86 setup.ini but not reported in cygwin-pkg-maint
>> at any level.
>> Nor the package, nor the source package and neither the first
>> subdirectory of x86/release is reported in the file;
>> I estimated the maintainer from the mailing list to be:
>>
>> PACKAGE                 SOURCE_PKG              MAINTAINER
>>
>> cloog-isl               cloog-isl               Achim Gratz
>> db4.5                   db4.5                   Dr. Volker Zell
>> db4.8                   db4.8                   Dr. Volker Zell
>> db4.8-doc               db4.8                   Dr. Volker Zell
>> font-misc-ethiopic      font-misc-ethiopic      Jon Turney
>> glib1.2                 glib1.2                 Yaakov Selkowitz
>> gtk1.2                  gtk1.2                  OBSOLETE (Yaakov
>> Selkowitz)
>> gtk1.2-engines          gtk1.2-engines          OBSOLETE (Yaakov
>> Selkowitz)
>
> The new upload mechanism doesn't seem to handle OBSOLETE well, so these
> two should be in my name without qualification in case I need to update
> them in the future.
>
>> html2ps                 html2ps                 Yaakov Selkowitz
>> isl                     isl                     Achim Gratz
>> libart_lgpl_2           libart_lgpl_2           Yaakov Selkowitz
>> libaudio-devel          nas                     Yaakov Selkowitz
>> libaudio2               nas                     Yaakov Selkowitz
>> libcloog-isl-devel      cloog-isl               Achim Gratz
>> libcloog-isl-doc        cloog-isl               Achim Gratz
>> libcloog-isl4           cloog-isl               Achim Gratz
>> libedit                 libedit                 Corinna Vinschen
>> libedit-devel           libedit                 Corinna Vinschen
>> libedit0                libedit                 Corinna Vinschen
>> libisl-devel            isl                     Achim Gratz
>> libisl-doc              isl                     Achim Gratz
>> libisl10                isl                     Achim Gratz
>> libzzip-devel           zziplib                 Yaakov Selkowitz
>> libzzip0.13             zziplib                 Yaakov Selkowitz
>> mesa-demos              mesa-demos              Yaakov Selkowitz
>> nas                     nas                     Yaakov Selkowitz
>> nc                      nc                      Corinna Vinschen
>> printproto              printproto              OBSOLETE (Yaakov
>> Selkowitz)
>> pwget                   pwget                   Jari Aalto
>> xhtml2ps                html2ps                 Yaakov Selkowitz
>> xlsx2csv                xlsx2csv                Jari Aalto
>> zziplib                 zziplib                 Yaakov Selkowitz
>
> The source packages should definitely be added; the subpackages depends
> on your last question.

OK.

>
>> 2) packages that are in cygwin-pkg-maint but not in setup.ini (nor x86
>> or x86_64 ones)
>>
>> most are OBSOLETE but not clearly defined as such, like
>> db2, db3.1, .. of Volker
>> libqhull, liblapack of mine
>
> Do you have a complete list?

yes, but looking at it I realized I need to merge 32bit and 64 bit 
analysis as some package exists only on 64bit due to the different splits.
I will report them later.


>
>> 3) typos
>>      libgck1-ddoc -> libgck1-doc
>>      libzziplib-devel -> libzzip-devel
>
> Yes, these are obvious fixes.
>
>> Question before I start to modify cygwin-pkg-maint:
>>
>> Does still make sense to report all binary packages in
>> the cygwin-pkg-maint files ?
>> Should we cluster them at source package or directory level as accepted
>> by the upload procedure ?
>> Specially for the dll libraries the name are continuously bumped, and
>> we will be always behind. At first glance we are missing ~ 200
>> packages; so roughly 8% of all the binary packages (~ 2700).
>> The active source packages are ~ 1400 and they change much
>> less frequently (ITP, ITA, orphanage and obsolescence)
>> so a better accuracy is possible.
>
> I've actually been thinking along the same lines.  AFAICS, the new
> upload mechanism doesn't care about subpackage listings so long as the
> source package is listed (provided, of course, that the standard
> directory layout is used).
>
> Therefore, the only "benefit" to having subpackages listed would be
> purely informative, but as you point out, 1) right now it's behind, 2)
> it's high maintenance for little gain, plus 3) a single person must
> anyway maintain an entire package with all its subpackages, and 4) nor
> does it actually indicate which subpackages map to which source packages.
>
> So I am all in favour of making this a source-only list, in line with
> other distro package maintainer "databases".
>
> Marco, if you're interested, please go ahead and add the missing
> *source* packages and fix those typos, but please hold off on removing
> anything until we hear from Corinna.

Ok, I will do.

> Thanks,
>
> Yaakov

Regards
Marco

Re: cygwin-pkg-maint maintance

[Corinna Vinschen]

[-- Attachment #1: Type: text/plain, Size: 2854 bytes --]

On Aug 10 23:55, Yaakov Selkowitz wrote:
> On 2014-08-10 15:33, Marco Atzeri wrote:
> >I was looking at the needed maintenance of cygwin-pkg-maint,
> >crossing the data of x86 setup.ini with cygwin-pkg-maint
> >and I found 2 main classes of mismatch plus a minor one
> 
> Thanks for taking the time to look into this!

Ditto!

> The source packages should definitely be added; the subpackages depends on
> your last question.
> 
> >2) packages that are in cygwin-pkg-maint but not in setup.ini (nor x86
> >or x86_64 ones)
> >
> >most are OBSOLETE but not clearly defined as such, like
> >db2, db3.1, .. of Volker
> >libqhull, liblapack of mine
> 
> Do you have a complete list?
> 
> >3) typos
> >     libgck1-ddoc -> libgck1-doc
> >     libzziplib-devel -> libzzip-devel
> 
> Yes, these are obvious fixes.
> 
> >Question before I start to modify cygwin-pkg-maint:
> >
> >Does still make sense to report all binary packages in
> >the cygwin-pkg-maint files ?
> >Should we cluster them at source package or directory level as accepted
> >by the upload procedure ?
> >Specially for the dll libraries the name are continuously bumped, and
> >we will be always behind. At first glance we are missing ~ 200
> >packages; so roughly 8% of all the binary packages (~ 2700).
> >The active source packages are ~ 1400 and they change much
> >less frequently (ITP, ITA, orphanage and obsolescence)
> >so a better accuracy is possible.
> 
> I've actually been thinking along the same lines.  AFAICS, the new upload
> mechanism doesn't care about subpackage listings so long as the source
> package is listed (provided, of course, that the standard directory layout
> is used).
> 
> Therefore, the only "benefit" to having subpackages listed would be purely
> informative, but as you point out, 1) right now it's behind, 2) it's high
> maintenance for little gain, plus 3) a single person must anyway maintain an
> entire package with all its subpackages, and 4) nor does it actually
> indicate which subpackages map to which source packages.
> 
> So I am all in favour of making this a source-only list, in line with other
> distro package maintainer "databases".
> 
> Marco, if you're interested, please go ahead and add the missing *source*
> packages and fix those typos, but please hold off on removing anything until
> we hear from Corinna.

No worries, this all makes sense to me.  The original list was based on
some automatism to fetch the info and then had to be kept in shape
manually and naturally deteriorated.

Let's keep the base (source) packages in the list and remove the
subpackages.

Thanks a lot for your help, Marco.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: cygwin-pkg-maint maintance

[Jon TURNEY]

On 10/08/2014 21:33, Marco Atzeri wrote:
> Hi Maintainers,

> PACKAGE                 SOURCE_PKG              MAINTAINER
>
> font-misc-ethiopic      font-misc-ethiopic      Jon Turney
I think I must decline this honour.  I guess that Yaakov was the 
uploader of this package.

Re: cygwin-pkg-maint maintance

[Marco Atzeri]

On 11/08/2014 13:20, Jon TURNEY wrote:
> On 10/08/2014 21:33, Marco Atzeri wrote:
>> Hi Maintainers,
>
>> PACKAGE                 SOURCE_PKG              MAINTAINER
>>
>> font-misc-ethiopic      font-misc-ethiopic      Jon Turney
> I think I must decline this honour.  I guess that Yaakov was the
> uploader of this package.
>

noted

Re: cygwin-pkg-maint maintance

[Marco Atzeri]

On 11/08/2014 12:44, Corinna Vinschen wrote:
> On Aug 10 23:55, Yaakov Selkowitz wrote:
>> On 2014-08-10 15:33, Marco Atzeri wrote:
>>> I was looking at the needed maintenance of cygwin-pkg-maint,
>>> crossing the data of x86 setup.ini with cygwin-pkg-maint
>>> and I found 2 main classes of mismatch plus a minor one
>>
>> Thanks for taking the time to look into this!
>
> Ditto!
>

>>
>> Marco, if you're interested, please go ahead and add the missing *source*
>> packages and fix those typos, but please hold off on removing anything until
>> we hear from Corinna.
>
> No worries, this all makes sense to me.  The original list was based on
> some automatism to fetch the info and then had to be kept in shape
> manually and naturally deteriorated.
>
> Let's keep the base (source) packages in the list and remove the
> subpackages.

Ok, I will prepare a revised cygwin-pkg-maint in the coming days
as soon I complete also the x86_64 analysis and merge all the data.

> Thanks a lot for your help, Marco.
>
>
> Corinna

You're welcome
Marco

Re: cygwin-pkg-maint maintance

[Marco Atzeri]

[-- Attachment #1: Type: text/plain, Size: 1573 bytes --]

On 11/08/2014 12:44, Corinna Vinschen wrote:
> On Aug 10 23:55, Yaakov Selkowitz wrote:
>> On 2014-08-10 15:33, Marco Atzeri wrote:
>>> I was looking at the needed maintenance of cygwin-pkg-maint,
>>> crossing the data of x86 setup.ini with cygwin-pkg-maint
>>> and I found 2 main classes of mismatch plus a minor one
>>
>> Thanks for taking the time to look into this!
>
> Ditto!
>

>>
>> Marco, if you're interested, please go ahead and add the missing *source*
>> packages and fix those typos, but please hold off on removing anything until
>> we hear from Corinna.
>
> No worries, this all makes sense to me.  The original list was based on
> some automatism to fetch the info and then had to be kept in shape
> manually and naturally deteriorated.
>
> Let's keep the base (source) packages in the list and remove the
> subpackages.
>
> Thanks a lot for your help, Marco.
>
>
> Corinna
>

attached 2 files.
The first is basically what should be the new cygwin-pkg-maint
that cover all the active package in both 32bit and 64 bit.
All packages are reported as lower case.

The second one is the list of packages reported in the current file
that do not exist as package or source package.
Except very few macro families (xorg, unison) all the rest seems to
me bumped dll's or old removed packages; I already removed some obvious 
typos, but I could have missed others.

Questions:
- Do we want to maintain the macro families in the files ?
- What to do with the other entries :
  leave as OBSOLETE record (at package level) or drop them completely ?

Regards
Marco






[-- Attachment #2: pkg-maint-active.txt.bz2 --]
[-- Type: application/octet-stream, Size: 9560 bytes --]

[-- Attachment #3: pkg-maint-obsolete.txt.bz2 --]
[-- Type: application/octet-stream, Size: 1010 bytes --]

Re: cygwin-pkg-maint maintance [gold star]

[Yaakov Selkowitz]

On 2014-08-11 17:03, Marco Atzeri wrote:
> attached 2 files.
> The first is basically what should be the new cygwin-pkg-maint
> that cover all the active package in both 32bit and 64 bit.

This was great, thank you.  After rearranging the release areas, I was 
able to get a more reliable list of what is currently on the server, and 
made a few more tweaks.

> All packages are reported as lower case.

That always annoyed be about this list, so I did change the list to use 
proper case.

> The second one is the list of packages reported in the current file
> that do not exist as package or source package.

We don't need those, so they're all gone, as well as those packages 
which are now nothing more than upgrade helpers.

Could you check the latest cygwin-pkg-maint and see if I missed anything?

Andrew: could you polish a gold star for Marco for showing initiative 
wrt Cygwin infrastructure at this juncture?

Thanks,

Yaakov

Re: cygwin-pkg-maint maintance [gold star]

[Marco Atzeri]

n 12/08/2014 07:22, Yaakov Selkowitz wrote:
> On 2014-08-11 17:03, Marco Atzeri wrote:
>
> That always annoyed be about this list, so I did change the list to use
> proper case.

I will reinstate proper case on next check.

>
> Could you check the latest cygwin-pkg-maint and see if I missed anything?

I will re-run possible today.

> Andrew: could you polish a gold star for Marco for showing initiative
> wrt Cygwin infrastructure at this juncture?
>
> Thanks,
>
> Yaakov

Regards
Marco

Re: cygwin-pkg-maint maintance [gold star]

[Marco Atzeri]

On 12/08/2014 13:34, Marco Atzeri wrote:
> n 12/08/2014 07:22, Yaakov Selkowitz wrote:
>> On 2014-08-11 17:03, Marco Atzeri wrote:
>>
>> That always annoyed be about this list, so I did change the list to use
>> proper case.
>
> I will reinstate proper case on next check.
>
>>
>> Could you check the latest cygwin-pkg-maint and see if I missed anything?
>
> I will re-run possible today.

No missing Maintainers for active packages (Category != _obsolete)

Minor issues on setup.hints for 32bit packages

   X-start-menu-icons
   libXaw8
   libXp6

they are now in the _obsolete tree but not declared as such in the 
setup.hint
As the last two have no source package they are jumping out as
anomalies on my script

I already cleaned a strange status of old postgresql libraries,
where some had a "[prev]" but not current lib.
All now has a proper empty current package and _obsolete category.

There are only two extra packages in the cygwin-pkg-maint,
that are absent in any form from the two setup.in

libxkbcommon                                 Yaakov Selkowitz
xcb-util-cursor                              Yaakov Selkowitz



>> Andrew: could you polish a gold star for Marco for showing initiative
>> wrt Cygwin infrastructure at this juncture?
>>

Regards
Marco

Re: cygwin-pkg-maint maintance [gold star]

[Yaakov Selkowitz]

On 2014-08-13 04:23, Marco Atzeri wrote:
> On 12/08/2014 13:34, Marco Atzeri wrote:
>> On 12/08/2014 07:22, Yaakov Selkowitz wrote:
>>> Could you check the latest cygwin-pkg-maint and see if I missed
>>> anything?
>>
>> I will re-run possible today.
>
> No missing Maintainers for active packages (Category != _obsolete)

Yeah!

> Minor issues on setup.hints for 32bit packages
>
>    X-start-menu-icons
>    libXaw8
>    libXp6
>
> they are now in the _obsolete tree but not declared as such in the
> setup.hint
> As the last two have no source package they are jumping out as
> anomalies on my script

They're probably ready for deletion, I just wasn't 100% sure yet.

> I already cleaned a strange status of old postgresql libraries,
> where some had a "[prev]" but not current lib.
> All now has a proper empty current package and _obsolete category.

Getting rid of all this cruft that nothing depends on anymore will be 
another project.

> There are only two extra packages in the cygwin-pkg-maint,
> that are absent in any form from the two setup.in
>
> libxkbcommon                                 Yaakov Selkowitz
> xcb-util-cursor                              Yaakov Selkowitz

I will be adding these shortly.

Thanks again for your help with this, it was greatly appreciated.


Yaakov

Re: cygwin-pkg-maint maintance [gold star]

[Andrew Schulman]

> On 2014-08-11 17:03, Marco Atzeri wrote:
> > attached 2 files.
> > The first is basically what should be the new cygwin-pkg-maint
> > that cover all the active package in both 32bit and 64 bit.
> 
> This was great, thank you.  After rearranging the release areas, I was 
> able to get a more reliable list of what is currently on the server, and 
> made a few more tweaks.
> 
> > All packages are reported as lower case.
> 
> That always annoyed be about this list, so I did change the list to use 
> proper case.
> 
> > The second one is the list of packages reported in the current file
> > that do not exist as package or source package.
> 
> We don't need those, so they're all gone, as well as those packages 
> which are now nothing more than upgrade helpers.
> 
> Could you check the latest cygwin-pkg-maint and see if I missed anything?
> 
> Andrew: could you polish a gold star for Marco for showing initiative 
> wrt Cygwin infrastructure at this juncture?

Awarded!  http://cygwin.com/goldstars/#MA

Re: cygwin-pkg-maint maintance

[Achim Gratz]

Yaakov Selkowitz writes:
[…]
> Yes, these are obvious fixes.
[…]

I've looked at the file today.  It seems that perl_vendor has been
removed, but not all Perl distributions that were bundled are listed.
I'll prepare a list of those over the weekend.

Given the purpose of the file and that it's meant to be consumed by
scripts, how about switching to something more easily parseable like
JSON or YAML (much like Perl does)?  That would enable to incorporate
more meta-information for each package which could easily be generated
by cygport.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#BlofeldUserWavetables

Re: cygwin-pkg-maint maintance

[Yaakov Selkowitz]

On Thu, 2014-08-14 at 19:55 +0200, Achim Gratz wrote:
> I've looked at the file today.  It seems that perl_vendor has been
> removed, but not all Perl distributions that were bundled are listed.
> I'll prepare a list of those over the weekend.

perl_vendor was a subpackage of the perl source package (it does not
have its own -src), so this is correct.  All perl-* packages currently
in either arch should be listed; if not, PTC.

As for whatever other perl module packages that need to be added in
order to replace perl_vendor, those will be added when we have an
approved update or ITA for them.

> Given the purpose of the file and that it's meant to be consumed by
> scripts, how about switching to something more easily parseable like
> JSON or YAML (much like Perl does)?  That would enable to incorporate
> more meta-information for each package which could easily be generated
> by cygport.

This is a package ownership database, not a package information
database.  What additional information do you think would be useful
here?


Yaakov

Re: cygwin-pkg-maint maintance

[Achim Gratz]

Yaakov Selkowitz writes:
> This is a package ownership database, not a package information
> database.  What additional information do you think would be useful
> here?

Whether the package is available for both architectures and if it's
already converted to cygport for instance.  From that database the
current file can easily be created if necessary and any manual changes
to the file could bootstrap a new entry in the database (so the new
maintainer can upload).


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#BlofeldUserWavetables

Re: cygwin-pkg-maint maintance

[Marco Atzeri]

On 14/08/2014 21:21, Achim Gratz wrote:
> Yaakov Selkowitz writes:
>> This is a package ownership database, not a package information
>> database.  What additional information do you think would be useful
>> here?
>
> Whether the package is available for both architectures

Wrong expectation.
It is in both architectures if it appears in both setup.ini;
any other solution will create duplicated information that finally
need alignment and it is error prone.

I plan to produce a list of sources by arch as by product of
the current analysis.

Please note that the two trees are not exactly equal so there are
packages available only in 64 and not in 32bit
(biber is the first in alphabetical order)


> and if it's
> already converted to cygport for instance.  From that database the
> current file can easily be created if necessary and any manual changes
> to the file could bootstrap a new entry in the database (so the new
> maintainer can upload).

The build methods is maintainer choice.
I use cygport but I don't see a reason to mandate it.

I do not follow the rest of your statement,
could you clarify the expected outcome ?

> Regards,
> Achim.

Regards
Marco

Re: cygwin-pkg-maint maintance

[Achim Gratz]

Marco Atzeri writes:
>> Whether the package is available for both architectures
>
> Wrong expectation.

So what?  I get how things are right now, that doesn't mean it has to
stay forever that way.

> It is in both architectures if it appears in both setup.ini;
> any other solution will create duplicated information that finally
> need alignment and it is error prone.

Ultimately the need for this file sholud go away except for
bootstrapping a new maintainer.

> I plan to produce a list of sources by arch as by product of
> the current analysis.
>
> Please note that the two trees are not exactly equal so there are
> packages available only in 64 and not in 32bit
> (biber is the first in alphabetical order)

I know, I've done similar analysis before (in Perl, if that matters).

> The build methods is maintainer choice.
> I use cygport but I don't see a reason to mandate it.

Corinna has already motioned in that direction and I'd say it makes good
sense to converge onto a single package building method.  That's the
only way to ever get an automatic build off the ground.  If and when
that happens the only thing a maintainer has to do is upload a new
cygport file and check the build logs before release.

> I do not follow the rest of your statement,
> could you clarify the expected outcome ?

Long term: You enter a new maintainer manually and when the package is
finally uploaded an extended setup.hint provides all the information
that goes into the package database to fill the temporary blanks.  And
with a little bit of extra effort each package could eben be checked for
(formal) correctness.


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

SD adaptation for Waldorf rackAttack V1.04R1:
http://Synth.Stromeko.net/Downloads.html#WaldorfSDada

Re: cygwin-pkg-maint maintance

[Corinna Vinschen]

[-- Attachment #1: Type: text/plain, Size: 1508 bytes --]

On Aug 14 22:28, Marco Atzeri wrote:
> On 14/08/2014 21:21, Achim Gratz wrote:
> >Yaakov Selkowitz writes:
> >>This is a package ownership database, not a package information
> >>database.  What additional information do you think would be useful
> >>here?
> >
> >Whether the package is available for both architectures
> 
> Wrong expectation.
> It is in both architectures if it appears in both setup.ini;
> any other solution will create duplicated information that finally
> need alignment and it is error prone.
> 
> I plan to produce a list of sources by arch as by product of
> the current analysis.
> 
> Please note that the two trees are not exactly equal so there are
> packages available only in 64 and not in 32bit
> (biber is the first in alphabetical order)
> 
> 
> >and if it's
> >already converted to cygport for instance.  From that database the
> >current file can easily be created if necessary and any manual changes
> >to the file could bootstrap a new entry in the database (so the new
> >maintainer can upload).
> 
> The build methods is maintainer choice.
> I use cygport but I don't see a reason to mandate it.

Tiny correction:  New packages should use cygport.  We should really all
use the same packaging system.  After all, you don't use rpm files on
Debian or dpkg files on Fedora.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: cygwin-pkg-maint maintance

[Marco Atzeri]

On 14/08/2014 22:51, Corinna Vinschen wrote:
> On Aug 14 22:28, Marco Atzeri wrote:

>> The build methods is maintainer choice.
>> I use cygport but I don't see a reason to mandate it.
>
> Tiny correction:  New packages should use cygport.  We should really all
> use the same packaging system.  After all, you don't use rpm files on
> Debian or dpkg files on Fedora.

"should" is not "must". ;-)

Of course it is the preferred and major way and
the alternatives are not so efficient.

> Corinna

Marco

Re: cygwin-pkg-maint maintance

[Corinna Vinschen]

[-- Attachment #1: Type: text/plain, Size: 801 bytes --]

On Aug 14 23:29, Marco Atzeri wrote:
> On 14/08/2014 22:51, Corinna Vinschen wrote:
> >On Aug 14 22:28, Marco Atzeri wrote:
> 
> >>The build methods is maintainer choice.
> >>I use cygport but I don't see a reason to mandate it.
> >
> >Tiny correction:  New packages should use cygport.  We should really all
> >use the same packaging system.  After all, you don't use rpm files on
> >Debian or dpkg files on Fedora.
> 
> "should" is not "must". ;-)

Yes, that's me being vague, as usual.  We're going to convert the
"should" to a "must" for new packages, though, and an additional
"should" for updates to older packages.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat

[-- Attachment #2: Type: application/pgp-signature, Size: 819 bytes --]

Re: cygwin-pkg-maint maintance: round 2

[Marco Atzeri]

[-- Attachment #1: Type: text/plain, Size: 1506 bytes --]

On 14/08/2014 22:47, Achim Gratz wrote:
> Marco Atzeri writes:

> Ultimately the need for this file should go away except for
> bootstrapping a new maintainer.

I guess we will always need a maintainer database

>> I plan to produce a list of sources by arch as by product of
>> the current analysis.

here it is including the crude python (*) script I am using:
http://matzeri.altervista.org/works/check_maint/

http://matzeri.altervista.org/works/check_maint/pkg_arch.txt

Attached the mismatch between 32 bit and 64 bit filtered by

$ grep -v -e "cygwin\|perl\|Both" pkg_arch.txt  |tee pkg_mismatch.txt

to remove the cross-package and the perl subpackages that
are the major part of the virtual mismatch.

They are around 100 but some are in reality obsoleted as
rxvt and aewm++ that are dead upstream, and some are
virtual mismatches.


> Long term: You enter a new maintainer manually and when the package is
> finally uploaded an extended setup.hint provides all the information
> that goes into the package database to fill the temporary blanks.  And
> with a little bit of extra effort each package could eben be checked for
> (formal) correctness.

big dream ;-)

For the time being I suggest to focus on the mismatches and see
which programs should be buried as obsolete and which are worth
to be ported on 64bit.


> Regards,
> Achim.

*) my real job is not in software, and writing a program every
    2-3 year  is a very good way to continuously loose practice
    and relearn a language...


[-- Attachment #2: pkg_mismatch.txt --]
[-- Type: text/plain, Size: 7363 bytes --]

PACKAGE                                      ARCH      MAINTAINER

aewm++                                       32 bit    ORPHANED (Jari Aalto)
aewm++-goodies                               32 bit    ORPHANED (Jari Aalto)
amanda                                       32 bit    Mark O'Keefe
apngtools                                    32 bit    Jari Aalto
biber                                        64 bit    Ken Brown
brltty                                       32 bit    ORPHANED (Samuel Thibault)
bsdiff                                       32 bit    ORPHANED (Lapo Luchini)
catdoc                                       32 bit    Reini Urban
ccdoc                                        32 bit    ORPHANED (Joe Linoff)
cfget                                        64 bit    Jari Aalto
cgoban                                       32 bit    Teun Burgers
chicken                                      32 bit    Nathan Thern
clisp                                        32 bit    Reini Urban
connect-proxy                                32 bit    Kostya Altukhov
cramfs                                       32 bit    ORPHANED (Sam Robb)
ctris                                        32 bit    Reini Urban
db                                           64 bit    Dr. Volker Zell
db4.5                                        32 bit    Dr. Volker Zell
db4.8                                        32 bit    Dr. Volker Zell
distcc                                       32 bit    ORPHANED (Harold L Hunt II)
dmalloc                                      32 bit    Pierre A. Humblet
exim                                         32 bit    Pierre A. Humblet
fcgi                                         32 bit    Reini Urban
ffcall                                       32 bit    Reini Urban
flexdll                                      32 bit    Yaakov Selkowitz
fltk_gdi                                     32 bit    Teun Burgers
gaffitter                                    32 bit    Kostya Altukhov
geoip                                        32 bit    Yaakov Selkowitz
git-oodiff                                   32 bit    Jari Aalto
gnucap                                       32 bit    Peter Li
gnugo                                        32 bit    Teun Burgers
grandr                                       32 bit    Yaakov Selkowitz
ioperm                                       32 bit    ORPHANED (Marcel Telka)
iperf                                        32 bit    Mike White
lesstif                                      32 bit    OBSOLETE (Yaakov Selkowitz)
libAfterImage                                32 bit    ORPHANED (Charles Wilson)
libassuan                                    32 bit    ORPHANED (Charles Wilson)
libgeotiff                                   32 bit    ORPHANED (Charles Wilson)
libksba                                      32 bit    ORPHANED (Charles Wilson)
libpthread-stubs                             32 bit    Yaakov Selkowitz
libsigsegv                                   32 bit    Reini Urban
libtextcat                                   32 bit    Reini Urban
libtorrent                                   32 bit    Chris Sutcliffe
lilypond                                     32 bit    Jan Nieuwenhuizen
logiweb                                      32 bit    Klaus Grue
lyx                                          32 bit    ORPHANED (Bo Peng)
maradns                                      32 bit    Steven Monai
mathomatic                                   32 bit    Reini Urban
md5deep                                      32 bit    Steven Monai
mingw-libgcrypt                              32 bit    ORPHANED (Charles Wilson)
mingw-libgpg-error                           32 bit    ORPHANED (Charles Wilson)
mingw-xz                                     32 bit    ORPHANED (Charles Wilson)
mtd                                          32 bit    ORPHANED (Sam Robb)
naim                                         32 bit    ORPHANED (Jonathan C. Allen)
nfrotz                                       32 bit    ORPHANED (Charles Wilson)
nfs-server                                   32 bit    ORPHANED (Sam Robb)
oodiff                                       32 bit    Jari Aalto
pal                                          32 bit    ORPHANED (Jari Aalto)
parrot                                       32 bit    Reini Urban
pbzip2                                       32 bit    Peter Li
pdftk                                        32 bit    David Rothenberger
pinentry                                     32 bit    ORPHANED (Charles Wilson)
ploticus                                     32 bit    Andrew Schulman
ploticus-doc                                 32 bit    Andrew Schulman
po4a                                         64 bit    Yaakov Selkowitz
printproto                                   32 bit    OBSOLETE (Yaakov Selkowitz)
pristine-tar                                 32 bit    Jari Aalto
proftpd                                      32 bit    Jason Tishler
proj                                         32 bit    ORPHANED (Charles Wilson)
pth                                          32 bit    ORPHANED (Charles Wilson)
pure-ftpd                                    32 bit    Kostya Altukhov
rakudo                                       32 bit    Reini Urban
rakudo-star                                  32 bit    Reini Urban
rtorrent                                     32 bit    Chris Sutcliffe
rxvt                                         32 bit    ORPHANED (Charles Wilson)
scsh                                         32 bit    Reini Urban
singular-base                                32 bit    Yue Ren (Frank Seelisch)
singular-help                                32 bit    Yue Ren (Frank Seelisch)
singular-icons                               32 bit    Yue Ren (Frank Seelisch)
singular-share                               32 bit    Yue Ren (Frank Seelisch)
singular-surf                                32 bit    Yue Ren (Frank Seelisch)
suite3270                                    32 bit    Peter A. Castro
sunrpc                                       32 bit    ORPHANED (Charles Wilson)
tack                                         32 bit    ORPHANED (Charles Wilson)
task                                         32 bit    Federico Hernandez
tesseract-ocr                                32 bit    Reini Urban
tinyirc                                      32 bit    Jari Aalto
ucspi-tcp                                    32 bit    Steven Monai
wtf                                          32 bit    Chris Sutcliffe
xclip                                        32 bit    Roland Cassard
xdelta                                       32 bit    Eric Blake
xemacs                                       32 bit    Dr. Volker Zell
xemacs-mule-sumo                             32 bit    Dr. Volker Zell
xemacs-sumo                                  32 bit    Dr. Volker Zell
xgraph                                       32 bit    ORPHANED (Jari Aalto)
xinetd                                       32 bit    ORPHANED (Charles Wilson)
xmltoman                                     64 bit    Yaakov Selkowitz
xsri                                         32 bit    ORPHANED (Charles Wilson)
xview                                        32 bit    Jon Turney
xwinwm                                       32 bit    Yaakov Selkowitz