Created attachment 274237 [details, diff] patch to chromium-12.0.742.60.ebuild Despite the assertion made by Paweł Hajdan, Jr. in Bug 357003, it's actually quite easy to build >=www-client/chromium-11 without needing PAM. $ equery list chromium pam * Searching for chromium ... [IP-] [ ] www-client/chromium-12.0.742.60:0 * Searching for pam ... !!! No installed packages matching 'pam'
I think a "remoting" use flag would be more appropriate than "pam".
(In reply to comment #1) > I think a "remoting" use flag would be more appropriate than "pam". "pam" is already a standard USE flag and will already be set according to whether the user wants that filth on their system or not.
(In reply to comment #2) USE flags are supposed to express optional features or functionality, not simply control dependencies. If pam integration was a feature of chromium, a pam use flag would make sense. Your patch is really disabling the remoting feature, which just happens to eliminate the pam dependency.
Is there a way to make the default value of the (local) "remoting" USE flag follow the currently effective value of the (global) "pam" USE flag? I don't know what the syntax for that would look like. Something like: IUSE="remoting[pam=]" I just looked through the Gentoo Development Guide, and I didn't see anything about having local USE flags default to the currently effective values of global USE flags. Oh well. I know I can always add "www-client/chromium -remoting" to my package.use. It just seems redundant if I already have USE="-pam" set system-wide.
(In reply to comment #4) > Is there a way to make the default value of the (local) "remoting" USE flag > follow the currently effective value of the (global) "pam" USE flag? > Not really. However, if we simply say IUSE="remoting", the flag will default to off. I don't think the feature is something most people would use, so this seems like a sensible thing to do anyway. IUSE="+remoting" would default it on.
It seems that chromium-13.x no longer links with PAM. Can anyone please confirm this? I tested with 13.0.767.1. (In reply to comment #0) > Despite the assertion made by Paweł Hajdan, Jr. in Bug 357003, it's actually > quite easy to build >=www-client/chromium-11 without needing PAM. I didn't say it is hard, but any optional dependency increases the maintenance effort. Each additional USE flag doubles the number of different build configurations. Please note that with other packages the ./configure flags usually are supported by upstream and are expected to work. This is not always the case with Chromium. To make it very clear: I'm *against* having an additional USE flag for the PAM dependency. However, I was never a fan of this solution (browser depending on PAM), and I hope it's clear from reading http://groups.google.com/a/chromium.org/group/chromium-dev/browse_thread/thread/3b7490cf60f56609 When we're sure that PAM is no longer needed (in the 13.x series), I'd be glad to remove that dependency.
(In reply to comment #6) > http://groups.google.com/a/chromium.org/group/chromium-dev/browse_thread/thread/3b7490cf60f56609 I did read that thread in its entirety before creating this patch. It is not clear from your comments that you disagree with disabling remoting. What is clear is that you disagree with a web browser having a dependency on PAM, and on that point I wholeheartedly agree with you. To me, the fact that "remoting" is a top-level build flag says that it's intended to be switchable. This isn't some intrusive and fragile hack to the build system; it's very simply flipping off a switch that is intended to be externally accessible.
On a more utilitarian note, it's a problem that there are no longer any versions of www-client/chromium in Portage that can be built without PAM. I had been sitting on 10.0.648.205 because it was the last version in Portage that didn't need PAM, but that version has now been removed, meaning it's now impossible for PAM-less Gentoo systems to emerge Chromium.
I just did a version bump for new dev channel release. >=chromium-13.0.772.0 (hard masked) doesn't have the PAM dependency. Feel free to try it. (In reply to comment #7) > To me, the fact that "remoting" is a top-level build flag says that it's > intended to be switchable. This isn't some intrusive and fragile hack to the > build system; it's very simply flipping off a switch that is intended to be > externally accessible. Those flags are frequently broken, really. The fact that it's there doesn't mean it's supported. It's more like "if it gets broken, we'll accept patches".
(In reply to comment #9) > I just did a version bump for new dev channel release. >=chromium-13.0.772.0 > (hard masked) doesn't have the PAM dependency. Feel free to try it. Thanks. I'm glad to see that Google has wised up. I'll stick with my hacked =chromium-12* until 13 gets unmasked.
(In reply to comment #10) > I'll stick with my hacked =chromium-12* until 13 gets unmasked. chromium-13.x is now in ~arch. Closing.
