headers
Stewart Smith | 1 Jun 2003 02:05
Picon

Re: upgrade php4 Portfile

I'm having a lot more success with +imap with this portfile (i.e. it's 
actually compiling now)

It gets my vote :)

On Saturday, May 31, 2003, at 08:29  PM, Benoit Chesneau wrote:

> Minor corrections and bump to new release of php4 : 4.3.2.
>
> http://www.opendarwin.org/bugzilla/show_bug.cgi?id=603
>
> Benoît Chesneau
> --
> Metavers,
> 30 rue Saint Sauveur - 75002 Paris (France)
> Tel/Fax 		:  +33 (0)1 4221 3909
> Tel (Mobile) 	:  +33 (0)6 7319 0113
> Web			: http://metavers.net
>
> _______________________________________________
> Darwinports mailing list
> Darwinports <at> opendarwin.org
> http://www.opendarwin.org/mailman/listinfo/darwinports
>
>
------------------------------
Stewart Smith
stewartsmith <at> mac.com
Ph: +61 4 3884 4332
ICQ: 6734154
(Continue reading)

Permalink | Reply | 0 comments |
headers
Stewart Smith | 1 Jun 2003 02:16
Picon

Re: Also about zsh

On Sunday, June 1, 2003, at 01:28  AM, Michael Maibaum wrote:
> It already can handle multiple distfiles, and lots of Portfiles already
> use this. I'd make a doc variant if the additional docs were considered
> particularly useful.

And an example is the imap variant in PHP. A copy&paste job could 
almost do the trick :)

------------------------------
Stewart Smith
stewartsmith <at> mac.com
Ph: +61 4 3884 4332
ICQ: 6734154
http://www.flamingspork.com/       http://www.linux.org.au
Permalink | Reply | 0 comments |
headers
Shantonu Sen | 1 Jun 2003 05:39

Progress report on RPM packaging

Ben and I made some good progress on this this week. We should have all 
the pieces of this ready. Creating binary RPM packages has 3 main 
parts:

1) RPM changes to support packaging 
(<http://www.opendarwin.org/bugzilla/show_bug.cgi?id=604>)
* We now have complete dependency checking that is way better than 
Installer.app (pffft). In addition to obvious explicit dependencies 
from the Portfile (we capture what ports a given port needs, and the 
versions at the time the RPM was built), we also support implicit 
shared library dependencies. RPM finds all the binaries in a package 
and grabs both what shared libraries/frameworks are required, and what 
are installed by the port
* We support fat packages. The shared library deps are really just 
string matching. So if a port depends on libbzip2.dylib, in the RPM 
this is really encoded as "/opt/local/lib/libbz2.dylib-ppc" (i.e. the 
install name of the library, suffixed with the architecture). If you 
were to build a fat bzip2 package, it would provice both "-ppc" and 
"-i386". In addition to just doing this shared lib dance, the RPM db 
itself knows whether a package is intended to be installed on multiple 
architectures. We made the "fat" architecture to be a superset of "ppc" 
and "i386", much like "i386" is a superset of "i686", since any 
unoptimized package can be installed on a PII. This is really more 
relevant to OpenDarwin users, since these patches will be integrated 
into the next OD release.

2) Implement RPM packaging 
(<http://www.opendarwin.org/bugzilla/show_bug.cgi?id=581>)
* The proposed patch makes a new target: "port rpmpackage foo", which 
identical semantics to "port package", except an RPM is generated
(Continue reading)

Permalink | Reply | 0 comments |
headers
Felix Kronlage | 1 Jun 2003 20:48

testing minicom

While Waqar has no problems building his submitted minicom port,
I seem to have problems compiling it (in a chroot and on a regular
install) , see the notes to bug #593.
Could anyone else test this and report wether it compiles or not?

-fkr
--

-- 
gpg-fingerprint: 076E 1E87 3E05 1C7F B1A0  8A48 0D31 9BD3 D9AC 74D0 
|Felix Kronlage                | whois -h whois.ripe.net FKR-RIPE |
|http://www.hazardous.org/fkr/ | fkr <at> {grummel.net|opendarwin.org} |
Permalink | Reply | 0 comments |
headers
Jordan Hubbard | 2 Jun 2003 00:02

Re: master_sites lists

I couldn't find any issues with it and I think it's a great 
improvement, so I just went ahead and committed it.  Good work!

- Jordan

On Thursday, May 29, 2003, at 03:21 AM, Will Barton wrote:

> Please note, that the biggest potential issue for Portfiles would be 
> that if you are using master_sites lists ([master-sites-for blah]), 
> then you MUST use master_sites.subdir AND that subdir will also be 
> appended to any sites you list in addition to the master_sites > list(s).
>
> Otherwise, I think this just needs to be tested and reviewed.  If 
> there aren't any issues brought up with this, I'll submit it to 
> bugzilla.
>
Permalink | Reply | 10 comments |
headers
Aristedes Maniatis | 2 Jun 2003 01:38
Picon
Favicon
Gravatar

Re: master_sites lists

Is there some documentation about how we should use this and which 
sites are already defined? I presume SourceForge will be one?

Ari Maniatis

On Monday, June 2, 2003, at 08:02  AM, Jordan Hubbard wrote:

> I couldn't find any issues with it and I think it's a great 
> improvement, so I just went ahead and committed it.  Good work!
>
> - Jordan
>
> On Thursday, May 29, 2003, at 03:21 AM, Will Barton wrote:
>
>> Please note, that the biggest potential issue for Portfiles would be 
>> that if you are using master_sites lists ([master-sites-for blah]), 
>> then you MUST use master_sites.subdir AND that subdir will also be 
>> appended to any sites you list in addition to the master_sites > 
>> list(s).
>>
>> Otherwise, I think this just needs to be tested and reviewed.  If 
>> there aren't any issues brought up with this, I'll submit it to 
>> bugzilla.
>>
>
> _______________________________________________
> Darwinports mailing list
> Darwinports <at> opendarwin.org
> http://www.opendarwin.org/mailman/listinfo/darwinports
>
(Continue reading)

Permalink | Reply | 9 comments |
headers
Benoit Chesneau | 2 Jun 2003 02:10

about apache2/apache rc script ....

Just a talk about the new activate_server variant in apache/apache 2 
dport. I think as I already told you that this is very confusing. In 
fact how somebody could activate the server if he hadn't installed the 
dport with this variant:  by now he should create its own rc script. I 
proposed to use something like rc-update (on gentoo) or chkconfig. But 
maybe there is another solution : why not using something similar to 
hostconfig in /etc on macos x ? If we have some similar file in 
${prefix}/etc/ it could solve the problem of activation, and we could 
use it in other rc file.

I think, that it could be very interresting to have a solution to 
activate/deactivate rc script. Sometimes we need indeed to do this on 
one script. What do you think about it ?

Benoît Chesneau
--
Metavers,
30 rue Saint Sauveur - 75002 Paris (France)
Tel/Fax 		:  +33 (0)1 4221 3909
Tel (Mobile) 	:  +33 (0)6 7319 0113
Web			: http://metavers.net
Permalink | Reply | 2 comments |
headers
Jordan Hubbard | 2 Jun 2003 02:46

Re: master_sites lists

On Sunday, June 1, 2003, at 04:38 PM, Aristedes Maniatis wrote:

> Is there some documentation about how we should use this and which 
> sites are already defined? I presume SourceForge will be one?

Documentation is definitely the next step.  I'll look at the relevant 
sections of the handbook / porter's guide if Will or Jim don't beat me 
to it.   Landon also said he has some stylistic issues with the 
implementation and I was going to give him a day or two to either 
decide they're not significant enough to do anything about, thus 
allowing us to simply document the mechanism as it stands, or to make 
whatever changes he feels would be more ascetically pleasing so we can 
document that.

- Jordan

>
> Ari Maniatis
>
>
> On Monday, June 2, 2003, at 08:02  AM, Jordan Hubbard wrote:
>
>> I couldn't find any issues with it and I think it's a great 
>> improvement, so I just went ahead and committed it.  Good work!
>>
>> - Jordan
>>
>> On Thursday, May 29, 2003, at 03:21 AM, Will Barton wrote:
>>
>>> Please note, that the biggest potential issue for Portfiles would be
(Continue reading)

Permalink | Reply | 8 comments |
headers
Jason Corley | 2 Jun 2003 02:53
Picon

Re: about apache2/apache rc script ....

I think the way to do this rather than relying on something like 
chkconfig symlink madness is to use xinetd as an example.  A simple line 
like:
    disable = yes
in each rc script would do the trick, and it would only require an extra 
bit of logic be added to DarwinPortsStartup to find the disable line and 
act on it accordingly.

Of course another idea I've proposed on IRC in the past is to use 
daemonic as the wrapper instead.  You could disable an item on startup 
with "sudo daemonic disable foo" and it would remove the SystemStarter 
bundle for you.

Jason

Benoit Chesneau wrote:

> Just a talk about the new activate_server variant in apache/apache 2 
> dport. I think as I already told you that this is very confusing. In 
> fact how somebody could activate the server if he hadn't installed the 
> dport with this variant:  by now he should create its own rc script. I 
> proposed to use something like rc-update (on gentoo) or chkconfig. But 
> maybe there is another solution : why not using something similar to 
> hostconfig in /etc on macos x ? If we have some similar file in 
> ${prefix}/etc/ it could solve the problem of activation, and we could 
> use it in other rc file.
>
> I think, that it could be very interresting to have a solution to 
> activate/deactivate rc script. Sometimes we need indeed to do this on 
> one script. What do you think about it ?
(Continue reading)

Permalink | Reply | 0 comments |
headers
Will Barton | 3 Jun 2003 08:30

Re: master_sites lists

I certainly don't mind documenting what I'v done, if the behavior 
doesn't change too much.  I can update the relevant sections of the 
Guide, but I'll wait on Landon's changes :)

-will

On Monday, June 2, 2003, at 02:46  AM, Jordan Hubbard wrote:

>
> On Sunday, June 1, 2003, at 04:38 PM, Aristedes Maniatis wrote:
>
>> Is there some documentation about how we should use this and which 
>> sites are already defined? I presume SourceForge will be one?
>
> Documentation is definitely the next step.  I'll look at the relevant 
> sections of the handbook / porter's guide if Will or Jim don't beat me 
> to it.   Landon also said he has some stylistic issues with the 
> implementation and I was going to give him a day or two to either 
> decide they're not significant enough to do anything about, thus 
> allowing us to simply document the mechanism as it stands, or to make 
> whatever changes he feels would be more ascetically pleasing so we can 
> document that.
>
> - Jordan
>
>>
>> Ari Maniatis
>>
>>
>> On Monday, June 2, 2003, at 08:02  AM, Jordan Hubbard wrote:
(Continue reading)

Permalink | Reply | 7 comments |

Gmane