In D25899#574151, @greg_unrelenting.technology wrote:

In D25899#574031, @rigoletto wrote:

I just still don't know why the cache was apparently being updated before but this is something I can investigate later.

Before as in..?

I just still don't know why the cache was apparently being updated before but this is something I can investigate later.

Add pkg-plist. In this particular case not using the pkg-plist file proved to be worse than using it with a few files.

Add PORTEPOCH. :-(

Well, I think this is more correct to be done at post-install phase. :-)

Seems to work as expected now.

In D25899#573804, @greg_unrelenting.technology wrote:

No, no, please read this carefully again:

Because the build system is doing gdk-pixbuf-query-loaders --update-cache for us, that should be disabled.
(And should be done in pkg post install hooks..)

We specifically MUST NOT run gdk-pixbuf-query-loaders --update-cache at BUILD time, only at pkg install time.

Basically copy the @postexec / @postunexec lines from graphics/librsvg2-rust/pkg-plist here.

General update.

The port was not abandoned good. :-)

Fix STRIP_CMD path.

Re-instate STRIP_CMD.

I think this is fine now. @uzsolt_uzsolt.hu is this good for you?

Sync with HEAD.
I removed all LD_CONFIG entries and it apparently is fine but I will still need to review it.

In D23908#565842, @uzsolt_uzsolt.hu wrote:

In D23908#565841, @rigoletto wrote:

@uzsolt_uzsolt.hu are you fine with it?

Sorry for late.
Why did you remove own mirror sites? Sometimes the pwmt.org is unavailable. If the source files (zathura-pdf-mupdf) are different I can update it on my host.

In D23908#565853, @mat wrote:

I do not think all the USE_LDCONFIG bits are needed.

The USE_LDCONFIG knob is there when the port provides a shared library that other ports will need to link with, here, the .so are plugins, and they will be dlopen'ed on request.

@uzsolt_uzsolt.hu are you fine with it?

In D23908#550315, @madpilot wrote:

I have no problems for the parts the I am maintainer of.

Make sure you have contacted the other maintainers and have permission or they are actually timed out.

Not all maintainers are subscribed to phabricator, so I don't think the review emails are enough to make sure maintainers are informed.

In D24851#547705, @tcberner wrote:

Good idea to split the conflicting part out.

Latest changes, including some portlint fixes.

In D23908#528010, @madpilot wrote:

In D23908#525749, @uzsolt_uzsolt.hu wrote:

In D23908#525724, @rigoletto wrote:

While working in some of the zathura-* ports the source files were missing (I can't recall which ones now) and that is why I discovered they moved to a GitLab instance. Then I assumed they will move completely at some point, and IMO this is a better idea to keep MASTER_SITES consistent among all zathura-* instead of be changing one by one when necessary.

I think would the best to use multiple MASTER_SITES.

I second this, it looks like a good idea, if the distfiles are the same.

WIP. Last works.

Major issue: disabling ENABLE_UTILS to some slaveports is not working at the moment.

In D23908#525723, @madpilot wrote:

Hi,

Why the change in MASTER_SITE? The project home page, when clicking on downloads still links to the https://pwmt.org/projects/girara/download/girara-0.3.4.tar.xz URL, so why not follow what upstream does?

I completely agree on removing the www.madpilot.net "mirror" which I added some time ago to fix a temporary problem.

NOT tested!