I'm requesting PearPC be added back to portage. I am attaching an ebuild for PearPC-0.5. I volunteer to maintain this package. Reproducible: Always
Created attachment 284077 [details] Ebuild to build and install PearPC 0.5
Created attachment 284079 [details, diff] Patch to prevent overriding CFLAGS
Good grief. First release in 5.5 years?! Hm, I might do this. Let me think about it.
I wouldn't say this depends on bug 320337; version 0.5 fixes the issues identified in that bug.
Created attachment 284145 [details, diff] Patch to change default interface selection As advised by Sebastian Biallas, the ebuild should default to the X interface rather than SDL. Dependencies are cleaned up a little, and an error message regarding USE flags is corrected.
pearpc was dropped
"Dropped" as in no maintainer stepped forward? I actually started this bug/request after PearPC was removed from Portage. My goal was to have it included once more.
(In reply to comment #7) > "Dropped" as in no maintainer stepped forward? > > I actually started this bug/request after PearPC was removed from Portage. My > goal was to have it included once more. I guess he meant "there is no such package so there is nothing to bump". Feel free to contact proxy-maint@ once you have something working
Created attachment 344308 [details] ebuild: EAPI patch * EAPI assignment in ebuild 'app-emulation/pearpc-0.5::x-portage' does * not conform with PMS section 7.3.1 (see bug #402167): * valid EAPI assignment must occur on or before line: 5
Created attachment 344310 [details, diff] ebuild: [patch] createdisk.py should be executable make /usr/share/pearpc-0.5/scripts/createdisk.py executable
Created attachment 344312 [details, diff] files/pearpc-0.5-amd64.patch wrt upstream http://sourceforge.net/tracker/?func=detail&aid=3387493&group_id=108675&atid=651195 NOTE: this patch is NOT very well tested and I'm NOT a coder, so I assume there ARE BUGS in it! What it should do: make --enable-cpu=jitc_x86_64 work as expected. I noticed, on my system if I add "amd64" it will select jitc_x86_64, otherwise it will use jitc_x86 (32-Bit) as described in the forementioned upstream bug.
Created attachment 344314 [details, diff] files/pearpc-0.5-cuda-parachute-enable.patch As mentioned in the Gentoo Forums: http://forums.gentoo.org/viewtopic-p-6970820-highlight-.html#6970820 This patch enables the lock-up parachute for the emulated CUDA IO. When I run PearPC with Mac OS X (any version, really) I get these errors and the emulation freezes: [IO/CUDA] <Warning> Event processing timed out. Event dropped. The lock-up parachute removed the freezes, thus the system (inside the emulation) will continue to be usable. Only the mouse position is occationally offset (it will no longer be able to move to one of the corners of the screen, but this is fixed by moving the mouse to the opposite corner once). Also, releasing the mouse from the emulator window (by pressing F12) does work as expected. After patching, the (error) message is: [IO/CUDA] <Warning> Event processing timed out. Event dropped. [IO/CUDA] <Warning> lock-up parachute
Created attachment 344316 [details, diff] ebuild: [patch] add amd64 and cuda-parachute epatches This adds the two patches amd64.patch and cuda-parachute-enable.patch from above to the ebuild.
Hello! I've found this bug and I'm happy that there acutally is an ebuild available. I added some patches in the comments above. My ARCH is amd64, and it seemed to build fine with the jit_x86_64 cpu selected, but when I run Mac OS X with it (Panther, Tiger) it displays the Apple logo and sits there forever. I don't know if it will work with another PPC OS, like BeOS or Linux/PPC. Anyway, with USE="-jit" (which is the default) it works. (And is terribly slow!) From the ebuild, I applied the interface selection patch, as GTK and Qt are broken at the moment. See upstream http://osdir.com/ml/emulators.pearpc.devel/2004-06/msg00775.html for example (although this would be for version 0.4, as this message is from 2004). I'm missing the files/settings file. What is in there? An example file would be great... Also I noticed that this settings file is installed to /usr/share/pearpc-0.5/settings, but the scripts that reside there (namely ifppc_up and ifppc_down) source this settings file from "./scripts/settings". So it would have to go into /usr/share/pearpc-0.5/scripts/settings, right? Last but not least, I'd be very happy to see this go into the official portage tree again. Yes, it is not ready for a production system, and yes, it is alpha status software. But still, an emulator such as this is a great peace of software! And aren’t there other unstable packages as well? It would still remain ~arch KEYWORDed -- so what?!? So, thanks for the ebuild and the patches so far!
After a cursory review, I didn't see anything wrong with your patches. I haven't rebuilt PearPC recently (not much to update), but I remember it including an example config file. I wasn't aware the if_up and if_down scripts referenced it, though I don't use them. I'll see how your patches run and look into the settings file in-depth later this week. Glad someone else is interested in PPC emulation on Gentoo. :) FYI, this bug sprang from the following forum thread: http://forums.gentoo.org/viewtopic-t-889582.html
Created attachment 344416 [details, diff] ebuild: [patch] settings insinto /usr/share/${P}/scripts File settings is sourced by /usr/share/${P}/ifppc_up and ifppc_down, so this is where it should be installed. NOTE: This file is still missing in my app-emulation/pearpc/files directory. Its contents could be reconstructed by exermining the above mentioned scripts. Comments on what is what are not so easy to reconstruct however. NOTE 2: The files are frome the pearpc source tarbal scripts directory, where the file in question, settings, is missing also.
Created attachment 520256 [details] ebuild: pearpc-0.6_pre20150712 The last version available is 0.6pre with patches from sebastianbiallas, as of this commit from 12 Jul 2015 at https://github.com/sebastianbiallas/pearpc/tree/7eb63de8a92b86e1b7438019ddcf935005112501. I've made an ebuild that pulls exactly that version. I've tested it with USE="X SDL -debug -jit". Changes from 0.5: 1. No more patches needed (they are already included) 2. In src_compile, ./autogen.sh was required. 3. I had to add "emake -j1" because with MAKEOPTS="-j2" or more, i.e. parallel build, it fails. (Failed for me with MAKEOPTS="-j16"...) Also I altered the installation directories a bit... It installs the scripts (and video.x) to /usr/share/pearpc-0.6/ because this is the version, even if it is a 0.6pre version. The doc directory however uses the ebuild version (0.6_pre20150712).
Created attachment 520258 [details] ebuild: pearpc-0.6_pre20150712 I've added a USE="extras" useflag which will add the extra downloads from http://pearpc.sourceforge.net/downloads.html (under "Extras"): 1. A sample 3 GB disk image (bzipped. 2. A sample 6 GB disk image (bzipped). 3. A network card driver ISO for Darwin/PPC and/or Mac OS X. The downloads are still available from the original sources. Also I made the scripts only install if USE="scripts" is used. This includes ifppc_up* and ifppc_down* as well as the createdisk.py python script.
Created attachment 520260 [details] ebuild: pearpc-0.6_pre20150712 And once more: fix permissions of /usr/share/pearpc-0.6/extras/* which were executable (0755) if USE="scripts" was also used (because it sets insopts for the scripts).
