qca-2 (beta2 to be exact) plugins.
Created attachment 89433 [details] qca-gnupg-0.1_p20060406.ebuild
Created attachment 89434 [details] qca-openssl-0.1_p20060406.ebuild
Created attachment 89435 [details] qca-pkcs11-0.1_p20060407.ebuild
Created attachment 89497 [details, diff] qca-gnupg_fix.diff fix gnupg crashing from http://el-tramo.be/files/psi/
Created attachment 89498 [details] qca-gnupg-0.1_p20060406.ebuild ebuild using patch provided
Added this to sunrise overlay. Add to patch pkcs11 to build with gcc 4
The "ugly workaround" inside qca-{gnupg,openssl,pkcs11}-*.ebuild is not only ugly, it's very broken. You must change the qca-{gnupg,openssl,pkcs11}.pro to disable debugging _before_ running qmake (or even configure?). This can be done with the following code: if ! built_with_use qt debug || ! use debug; then sed -i -e 's/CONFIG += debug/#CONFIG += debug/g' ${PN}.pro fi; Otherwise psi (and other apps using qca) will fail because QT and/or app/psi has debugging disabled and the qca-libs have it enabled. Changing the options insode the Makefile won't change anything but avoid the error.
(In reply to comment #7) > Otherwise psi (and other apps using qca) will fail because QT and/or app/psi > has debugging disabled and the qca-libs have it enabled. Changing the options > insode the Makefile won't change anything but avoid the error. thx for the tip! I wasn't aware about this. Anyway - qmake is sometimes "evil" :/ When I started using it, I found it very usefull. Now I'm starting to change my opinion. In every qt app, devs can put debug into pro file, or not. So to package it, we would have to put into every ebuild a patch or some code to disable debug :/ It would be best if qmake could use something like "CONFIG -= debug"... Have to read some more docs at trolltech. Maybe it can be done like this... I'll take care of this in few days. Just need some time to read and think. Best regards, Przemek
(In reply to comment #8) > In every qt app, devs can put > debug into pro file, or not. I think thats exactly the idea behind this. The devs put "CONFIG += debug" into the pro-File while they are debugging their software. If they build a release out of their dev-tree they simply change this to "CONFIG += release" and everything works fine. So it's mainly a dev-thing to create clean releases I think. The current version of all qca-apps are not stable, so they did not spend any time in changing this. But the final release will have this fixed I think. Btw.: This is the same thing you can do in good old Makefiles. Define a variable CONFIG that is added to every gcc-call and use some debug-defines inside your programm. (#ifdef DEBUG inside the c-files and optional -DDEBUG inside the Makefile)
> I think thats exactly the idea behind this. The devs put "CONFIG += debug" into > the pro-File while they are debugging their software. If they build a release > out of their dev-tree they simply change this to "CONFIG += release" and > everything works fine. Yes, but this isn't a friendly solution for packagers. Few qt apps I made ebuilds for, have non-standard CONFIG vars. For a global solution, defining own CONFIG isn't a good idea. We can make some "super-sed" rules for this, but this isn't a pernament solution also :/ I'm starting to fill, that even I don't fill like, seding would be best, it will be best what we can get.
I'm going to maintain qca, as Gustavo have no time for it now. In few days I'm going to add ebuilds into portage, but for now without any valid keyword (which means : only for testers who knows what they are doing). Regards, Przemek
libqca-gnupg, qca-openssl, compiles on amd64 (qca-pkcs11 need some patches for gcc-4 and amd64)
Created attachment 107767 [details, diff] qca-pkcs11-gcc4.patch can some one confirm gcc 4 problem (and solution) qca-pkcs11 compiled on my amd64
(In reply to comment #13) > Created an attachment (id=107767) [edit] > solve my GCC-4 problem > > can some one confirm gcc 4 problem (and solution) > > qca-pkcs11 compiled on my amd64 I didn't check it, but it looks like a valid fix for some gcc4 issues ;) Cheers, Przemek
looks like a valid gcc4 fix to me ;) Cheers, Przemek
Yes. It was my fault :( Sorry.
Created attachment 119933 [details] qca-openssl-0.1_p20070508.ebuild Version bump for qca-openssl plugin - needed for psi-0.11_rc1 without bundled qca I'm not sure are rest of plugins really need for now. they all are included in kde4... Cheers, Przemek
Created attachment 119939 [details, diff] qca-openssl-slotted-qca2.patch patch for slotted qca Cheers, Przemek
pls add amd64 keyword
Created attachment 120088 [details] qca-openssl-0.1_p20070508.ebuild Works with new slotting from bug#129536.
Created attachment 120089 [details] qca-pkcs11-0.1_p20070508.ebuild Same here.
Created attachment 120094 [details] qca-gnupg-0.1_p20070510.ebuild
Created attachment 120096 [details] qca-gnupg-0.1_p20070510.ebuild
Created attachment 120099 [details] qca-cyrus-sasl-0.1_p20070508.ebuild
Created attachment 120201 [details] qca-openssl-0.1_p20070508.ebuild Added LD_LIBRARY_PATH to configure so it find the right QCA.
Created attachment 120202 [details] qca-pkcs11-0.1_p20070508.ebuild Added LD_LIBRARY_PATH to configure so it find the right QCA.
Created attachment 120204 [details] qca-gnupg-0.1_p20070510.ebuild Added LD_LIBRARY_PATH to configure so it find the right QCA.
Created attachment 120205 [details] qca-cyrus-sasl-0.1_p20070508.ebuild Added LD_LIBRARY_PATH to configure so it find the right QCA.
qca-openssl-0.1_p20070508 ebuild from comment #25 builds and works fine here. Compiled against qca-2.0.0_beta5 from https://bugs.gentoo.org/show_bug.cgi?id=129536#c30. I'm using psi-0.11_rc1 with SSL enabled to test it.
No point in managing this on separate bug. *** This bug has been marked as a duplicate of bug 129536 ***