- User Since
- Apr 10 2016, 5:00 PM (183 w, 6 d)
Thu, Oct 17
Wed, Oct 9
Fri, Oct 4
It looks like your print/freetype2 is configured with the PNG knob on, so PNG is being pulled in and x11/xlockmore is linking to PNG when it is discovered. Fortunately xlockmore has configure options for both freetype2 and png (see below), so we could make this deterministic. I suspect many users including myself just use xlockmore to lock their screens (I just use the blank screensaver and immediately turn off the monitor), so things like freetype2 and png might be more appropriate as OFF by default. We could also turn those options on when something like MESAGL and/or MAGICK are turned. Or, we could create an BELLS_AND_WHISTLES_FANCY_SCREESAVERS knob (obvious not that name) and pull all these things in together. What do you think?
print/freetype2 does not have a hard dependency on graphics/png unless you turn on the non-default PNG knob.
It's also not the version change. Here are the build logs (default options) with 5.59:
Those are not identical builds to the patch that I submitted, are they? I see the version is 5.59, so I can't really say.
I didn't get any errors with poudriere testport building in 11/12 i386/amd64 with the default options.
Mon, Sep 30
I can imagine users who prefer one kerberos implementation would be frustrated to have to install a different implementation just for x11/xlockmore. I thought adding the other kerboros options, even though the build currently fails with them, would indicate that we are not intentionally forcing them to use one specific kerberos implementation. I don't have a strong opinion either way, so I returned it back to they way you prefer, one option for kerberos5 support. Does this look okay?
Fri, Sep 27
Wed, Sep 25
Hi Dave; is this ready to go after adding aarch64?
Sat, Sep 21
Sep 19 2019
- Add radio options for different Kerberos5 implementations. Only MIT and NONE are currently building.
- Remove --disable-setuid from CONFIGURE_ARGS because both xlock and play.sh are installed with setuid bits on.
Re-add GTK2 and KERBEROS5 options based on https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=240526
Sep 18 2019
Sep 16 2019
Sep 11 2019
I made those comments before I was completely clear on what was going on. MB was never really disable before the switch to option helpers. The fonts look like terminal fonts without it, so I think most users would prefer it on. That's why my fonts changed after this latest update even though I had never explicitly turned on the MB knob. It's a similar situation with the Xinerama knob. It wasn't being disabled before and that's why I thought that the login screen was suddenly not being displayed on the primary display after the latest update. Later I realized what was really going on. The login screen was expanded over all displays without Xinerama on.
Sep 10 2019
Sep 9 2019
Sep 6 2019
This is a reasonable approach, however I don't see the value in defining EMACS_REV since we are essentially doing VAR1=some_value then the only reference to VAR1 is VAR2=VAR1. I sort of feel the same about EMACS_VER, since it holds the same value as DISTVERSION in editors/emacs and almost the same value in editors/emacs-devel. It seems to add an extra layer (What's the value of EMACS_VER? It's DISTVERSION. What's the value of DISTVERSION? It's 26.3).
Sep 5 2019
Sep 4 2019
Also remove EMACS_VER since it's only used once
Sep 1 2019
Aug 30 2019
Aug 27 2019
Aug 26 2019
Aug 24 2019
Aug 17 2019
dch@, since you are away, I think I will pull the trigger on this. I'm fairly comfortable that the latest changes are improvements. Let's get this out to users and if you have changes you want to make, you can always make them when you return.
Add pkg-message to let users know that git-cola wants devel/git to be built with the GUI option on
Aug 16 2019
devel/git-cola: Update to v3.4; Add NLS knob; patch link creation
Aug 13 2019
Aug 9 2019
Aug 8 2019
editors/emacs: [WIP] re-add canna information to distinfo
Aug 6 2019
Aug 5 2019
Aug 4 2019
Aug 1 2019
Jul 29 2019
No problem. I'll commit it within a few days.
Jul 28 2019
Jul 20 2019
Jul 15 2019
Jul 1 2019
Jun 29 2019
Jun 28 2019
If you do want users to get a warning, you could bump PORTREVSION *AND* add a pkg-message.
I should clarify. Everything looks except the message about the bad symlink that Jonathan already mentioned.
Here are the changes I made against above. There is probably a better way to share this diff, but I'm short on time.
The poudriere testport matrix with the changes I described above all look good.
Jun 27 2019
If I build in poudriere, I see pkg-plist errors related to the language files.
Ooops.. I did not mean to accept this before addressing @olgeni's suggestion about the dependency.
I was referring to the PORTREVISION bump on lang/erlang-runtime.
Jun 24 2019
Jun 21 2019
Jun 16 2019
Jun 10 2019
I will test 11/12 i386/amd64 as soon as our build box becomes available, which will hopefully be tomorrow morning (local time).
Jun 8 2019
Jun 1 2019
May 30 2019
May 29 2019
I don't think users will get a warning when installing the package. I also don't think there is any change in the package, so bumping PORTREVISION is probably not necessary.
I don't think users will get a warning when installing the package. I also don't think there is any change in the package, so bumping PORTREVISION is probably not necessary. @olgeni's suggestion?
May 28 2019
May 26 2019
Oops, I somehow sent that early. I was going to say...