Re: Why do so many builds fail?

[Brandon Allbery]

On Wed, Mar 9, 2016 at 11:24 PM, Bachsau  wrote:

> It is just somewhat frustrating when you're coding and get in need for
> some librarys, then do a port install or port upgrade outdated and it
> breaks the whole thing.


OS X is probably not the platform you want to use if you're doing this
often. Likewise any other rolling upgrade platform (*BSD ports, Linux
Arch/Gentoo/derivatives thereof, Debian unstable or testing, etc.). (Fink
doesn't roll quite as often as MacPorts and Homebrew, but it's still not a
stable development platform; it's more like Debian testing.)

-- 
brandon s allbery kf8nh   sine nomine associates
allber...@gmail.com  ballb...@sinenomine.net
unix, openafs, kerberos, infrastructure, xmonadhttp://sinenomine.net
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users

Re: Why do so many builds fail?

[Ryan Schmidt]

> On Mar 9, 2016, at 10:24 PM, Bachsau  wrote:
> 
> 
>> Am 09.03.2016 um 17:01 schrieb Ryan Schmidt :
>> 
>> There is no "stable" release of MacPorts. Everything is provided as-is and 
>> best-effort. We are all volunteers, not being paid to work on this.
> 
> I know, thats okay. I don't demand anyone to support his ports like a 
> commercial product and no one is forced to supply always the latest version, 
> thouroughly tested. I'm also a developer and contribute to others OSS 
> projects or my own in my spare time. When I release something to the public I 
> also don't give any guarantees. But, when I push out something that is 
> supposed to work for many people, I try to make it work, because my 
> motivation for sharing it is to make people happy, not causing trouble, or 
> else I'm not going to release it or share it in ways where people don't 
> expect it to work out of the box, a git repository for example. If somethings 
> not ready or you don't have time to test it or even continue development, 
> thats perfectly fine, and I don't demand getting all the latest bleeding-edge 
> versions in time. No one is forced to contribute to MacPorts, but if you do, 
> it should work.
> 
> It is just somewhat frustrating when you're coding and get in need for some 
> librarys, then do a port install or port upgrade outdated and it breaks the 
> whole thing. People then often tell me: If you don't like it, fix it 
> yourself. But I can't contribute to everything I use, so there's projects I 
> contribute to and projects I don't. If you don't know about the codebase of a 
> project its hard and time consuming to get in. Furthermore, I don't know that 
> much about C/C++. Enough to do some slight modifications but when something 
> horribly fails, especialy in build phase, I'm having a very hard time fixing 
> this myself. This is why I report it to mailing lists or bug trackers, as for 
> a maintainer it's often much easier to understand whats going on. I know 
> you're all volunteers. Sorry for being that pissy yesterday.

I understand everything you've said. Don't really have much more to add. 
Obviously nobody is going to commit something they believe is broken, but it 
does sometimes end up being the case for some subset of users. When it does, 
and we learn that it has happened, we try to fix it. If everybody had to test 
everything on a clean system on every version of OS X and every version of 
Xcode before committing, nobody would ever commit anything because nobody has 
the time and resources to do all that testing. We do have automated build 
machines that do build every commit on a clean system with no ports installed 
with several versions of OS X and the latest version of Xcode for those 
systems. That automated system did catch this webkit2-gtk build problem on El 
Capitan, and the maintainer was able to fix it. It did take some time, because 
the maintainer wanted to verify the fix worked before committing it, and 
webkit2-gtk takes hours to build.


>> Am 09.03.2016 um 18:12 schrieb David Evans :
>> 
>> Was this on El Capitan or another version of OS X?  Can you attach (to the 
>> ticket) a copy of your build log (compressed)
>> from a clean build showing the failure mode?  What variants are you using?  
>> Anything else relevant?
> 
> I just tried to upgrade again and it succeeded. Glad to see you got it fixed, 
> but I will consider this in future reports. Was on El Capitan and variants 
> are -x11 +no_x11 +quartz +bash_completion.
> 
> PS: Is there a way to tell port upgrade to just skip failed builds and 
> continue to upgrade the rest of the packages? That could come in handy.

Yes, to an extent. This question comes up often.

sudo port upgrade outdated and not webkit2-gtk

However, if any other outdated port (let's say yelp for example) depends on 
webkit2-gtk, attempting to upgrade that port will again try to upgrade 
webkit2-gtk. If you don't want that, you have to list that other port too.

sudo port upgrade outdated and not \( webkit2-gtk yelp \)

And so on.

___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users

Re: Why do so many builds fail?

[Bachsau]

> Am 09.03.2016 um 17:01 schrieb Ryan Schmidt :
> 
> There is no "stable" release of MacPorts. Everything is provided as-is and 
> best-effort. We are all volunteers, not being paid to work on this.

I know, thats okay. I don't demand anyone to support his ports like a 
commercial product and no one is forced to supply always the latest version, 
thouroughly tested. I'm also a developer and contribute to others OSS projects 
or my own in my spare time. When I release something to the public I also don't 
give any guarantees. But, when I push out something that is supposed to work 
for many people, I try to make it work, because my motivation for sharing it is 
to make people happy, not causing trouble, or else I'm not going to release it 
or share it in ways where people don't expect it to work out of the box, a git 
repository for example. If somethings not ready or you don't have time to test 
it or even continue development, thats perfectly fine, and I don't demand 
getting all the latest bleeding-edge versions in time. No one is forced to 
contribute to MacPorts, but if you do, it should work.

It is just somewhat frustrating when you're coding and get in need for some 
librarys, then do a port install or port upgrade outdated and it breaks the 
whole thing. People then often tell me: If you don't like it, fix it yourself. 
But I can't contribute to everything I use, so there's projects I contribute to 
and projects I don't. If you don't know about the codebase of a project its 
hard and time consuming to get in. Furthermore, I don't know that much about 
C/C++. Enough to do some slight modifications but when something horribly 
fails, especialy in build phase, I'm having a very hard time fixing this 
myself. This is why I report it to mailing lists or bug trackers, as for a 
maintainer it's often much easier to understand whats going on. I know you're 
all volunteers. Sorry for being that pissy yesterday.

> Am 09.03.2016 um 18:12 schrieb David Evans :
> 
> Was this on El Capitan or another version of OS X?  Can you attach (to the 
> ticket) a copy of your build log (compressed)
> from a clean build showing the failure mode?  What variants are you using?  
> Anything else relevant?

I just tried to upgrade again and it succeeded. Glad to see you got it fixed, 
but I will consider this in future reports. Was on El Capitan and variants are 
-x11 +no_x11 +quartz +bash_completion.

PS: Is there a way to tell port upgrade to just skip failed builds and continue 
to upgrade the rest of the packages? That could come in handy.

smime.p7s
Description: S/MIME cryptographic signature
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users

Re: Why do so many builds fail?

[David Evans]

On 3/9/16 8:52 AM, Gustaf Neumann wrote:
> Am 09.03.16 um 16:59 schrieb David Evans:
>> On 3/9/16 12:05 AM, Bachsau wrote:
>>> Today I was running an upgrade and it failed building webkit2-gtk and I'm 
>>> somewhat pissy about this, as I don't want
>>> to be forced debugging ports before I can continue with my coding. Is there 
>>> any stable release of macports or and how
>>> comes that buggy ports make it into the ports tree all the time? Doesn't 
>>> anyone ever test his creations before
>>> publishing them?
>> An update was committed last evening (r146463) which seemed to work on 
>> initial testing but now is failing on El Capitan
>> only AFAICT.
> for what it is worth: webkit2-gtk was rebuild on my 10.11.3 notebook today 
> successfully.
> 
> all the best
> -g
> 
> ~% sudo port info -v webkit2-gtk
> webkit2-gtk @2.11.91_1 (www, gnome)
> Variants: debug, [+]gtk2, minibrowser, quartz, universal, [+]x11
> 
> Description:  Apple's WebKit2 HTML rendering library for GTK+3 (with 
> optional support for GTK+2 plugins)
> Homepage: http://webkitgtk.org/
> 
> Extract Dependencies: xz
> Build Dependencies:   cmake, gperf, gtk-doc, pkgconfig, python27, 
> py27-simplejson
> Library Dependencies: atk, at-spi2-atk, bison, cairo, flex, fontconfig, 
> freetype, enchant, geoclue2, glib2,
> gobject-introspection, gstreamer1, gstreamer1-gst-plugins-bad, 
> gstreamer1-gst-plugins-base, gstreamer1-gst-plugins-good,
> gtk3, harfbuzz-icu, hyphen, icu, libnotify, libpng, libsecret, libsoup, 
> libxml2, libxslt,
>   sqlite3, webp, zlib, mesa, xorg-libXt, gtk2
> Conflicts with:   webkit2-gtk-devel
> Platforms:darwin, freebsd
> License:  LGPL-2+ BSD
> Maintainers:  jerem...@macports.org, dev...@macports.org
> 
> ~% sudo port installed webkit2-gtk
> The following ports are currently installed:
>   webkit2-gtk @2.11.3_1+gtk2+llvm37+video
>   webkit2-gtk @2.11.90_0+gtk2+x11
>   webkit2-gtk @2.11.91_0+gtk2+x11
>   webkit2-gtk @2.11.91_1+gtk2+x11 (active)
> 

Thanks for the additional data point.  It did fail, however, on the El Capitan 
buildbot last night so something's going
on.  I'm doing a clean build now to see if I can reproduce the issue.

It would help if Bachsau would comment on the details of his failure.  
Actually, best thing is to file a ticket with,
at least, the following information:

Was this on El Capitan or another version of OS X?  Can you attach (to the 
ticket) a copy of your build log (compressed)
from a clean build showing the failure mode?  What variants are you using?  
Anything else relevant?

Thanks, Dave


___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users

Re: Why do so many builds fail?

[Gustaf Neumann]

Am 09.03.16 um 16:59 schrieb David Evans:

On 3/9/16 12:05 AM, Bachsau wrote:

Today I was running an upgrade and it failed building webkit2-gtk and I'm 
somewhat pissy about this, as I don't want to be forced debugging ports before 
I can continue with my coding. Is there any stable release of macports or and 
how comes that buggy ports make it into the ports tree all the time? Doesn't 
anyone ever test his creations before publishing them?

An update was committed last evening (r146463) which seemed to work on initial 
testing but now is failing on El Capitan
only AFAICT.
for what it is worth: webkit2-gtk was rebuild on my 10.11.3 notebook 
today successfully.


all the best
-g

~% sudo port info -v webkit2-gtk
webkit2-gtk @2.11.91_1 (www, gnome)
Variants: debug, [+]gtk2, minibrowser, quartz, universal, [+]x11

Description:  Apple's WebKit2 HTML rendering library for GTK+3 
(with optional support for GTK+2 plugins)

Homepage: http://webkitgtk.org/

Extract Dependencies: xz
Build Dependencies:   cmake, gperf, gtk-doc, pkgconfig, python27, 
py27-simplejson
Library Dependencies: atk, at-spi2-atk, bison, cairo, flex, fontconfig, 
freetype, enchant, geoclue2, glib2, gobject-introspection, gstreamer1, 
gstreamer1-gst-plugins-bad, gstreamer1-gst-plugins-base, 
gstreamer1-gst-plugins-good, gtk3, harfbuzz-icu, hyphen, icu, libnotify, 
libpng, libsecret, libsoup, libxml2, libxslt,

  sqlite3, webp, zlib, mesa, xorg-libXt, gtk2
Conflicts with:   webkit2-gtk-devel
Platforms:darwin, freebsd
License:  LGPL-2+ BSD
Maintainers:  jerem...@macports.org, dev...@macports.org

~% sudo port installed webkit2-gtk
The following ports are currently installed:
  webkit2-gtk @2.11.3_1+gtk2+llvm37+video
  webkit2-gtk @2.11.90_0+gtk2+x11
  webkit2-gtk @2.11.91_0+gtk2+x11
  webkit2-gtk @2.11.91_1+gtk2+x11 (active)

___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users

Re: Why do so many builds fail?

[Ryan Schmidt]

On Mar 9, 2016, at 2:05 AM, Bachsau  wrote:

> Today I was running an upgrade and it failed building webkit2-gtk and I'm 
> somewhat pissy about this, as I don't want to be forced debugging ports 
> before I can continue with my coding. Is there any stable release of macports 
> or and how comes that buggy ports make it into the ports tree all the time? 
> Doesn't anyone ever test his creations before publishing them?

Sorry for the trouble. Of course we want your experience with MacPorts to be 
trouble-free, but we cannot guarantee that it will be, and problems do occur 
from time to time. When they do, please submit a bug report in the issue 
tracker.

Yes, everyone tests their changes before committing. But they probably only 
test on their own system. It may be that a build fails only on certain versions 
of OS X, or certain versions of Xcode, or only when certain other ports were 
already installed. There are countless possible reasons for a build failure. In 
each case, we must diagnose the issue and fix it.

There is no "stable" release of MacPorts. Everything is provided as-is and 
best-effort. We are all volunteers, not being paid to work on this.

___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users

Re: Why do so many builds fail?

[David Evans]

On 3/9/16 12:05 AM, Bachsau wrote:
> Today I was running an upgrade and it failed building webkit2-gtk and I'm 
> somewhat pissy about this, as I don't want to be forced debugging ports 
> before I can continue with my coding. Is there any stable release of macports 
> or and how comes that buggy ports make it into the ports tree all the time? 
> Doesn't anyone ever test his creations before publishing them?

An update was committed last evening (r146463) which seemed to work on initial 
testing but now is failing on El Capitan
only AFAICT.  Builds OK on Yosemite and Mavericks according to the buildbots.  
We (the port maintainers) are working on
the issue and hope to have it fixed shortly.

Please accept our apologies for the inconvenience.
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users

Why do so many builds fail?

[Bachsau]

Today I was running an upgrade and it failed building webkit2-gtk and I'm 
somewhat pissy about this, as I don't want to be forced debugging ports before 
I can continue with my coding. Is there any stable release of macports or and 
how comes that buggy ports make it into the ports tree all the time? Doesn't 
anyone ever test his creations before publishing them?

smime.p7s
Description: S/MIME cryptographic signature
___
macports-users mailing list
macports-users@lists.macosforge.org
https://lists.macosforge.org/mailman/listinfo/macports-users