It's needed since 0.5.9 is not longer supported upstream and it isn't compatible with gtk+-2.18 Reproducible: Always
Created attachment 219491 [details] grisbi-0.6.0_rc2.ebuild This is based on zugaina ebuild for it: http://gpo.zugaina.org/AJAX/Ebuild/75578/View But take care that grisbi won't work (at least for me) with cunit USE flag enabled: $ grisbi CUnit - A Unit testing framework for C - Version 2.1-0 http://cunit.sourceforge.net/ (process:29254): GLib-CRITICAL **: g_date_compare: assertion `rhs != NULL' failed (process:29254): GLib-CRITICAL **: g_date_compare: assertion `rhs != NULL' failed (process:29254): GLib-CRITICAL **: g_date_compare: assertion `rhs != NULL' failed (process:29254): GLib-CRITICAL **: g_date_compare: assertion `rhs != NULL' failed (process:29254): GLib-CRITICAL **: g_date_compare: assertion `rhs != NULL' failed (process:29254): GLib-CRITICAL **: g_date_compare: assertion `rhs != NULL' failed Suite gsb_data_account, Test of gsb_data_account() had failures: 1. gsb_data_account_cunit.c:90 - CU_ASSERT_EQUAL(0x80000000,balance.mantissa) 2. gsb_data_account_cunit.c:91 - CU_ASSERT_EQUAL(0,balance.exponent) Suite gsb_real, Test of gsb_real_get_from_string() had failures: 1. gsb_real_cunit.c:74 - CU_ASSERT_EQUAL(12345,val.mantissa) 2. gsb_real_cunit.c:75 - CU_ASSERT_EQUAL(2,val.exponent)Violación de segmento (Violacion de segmento = Segmentation Fault)
Created attachment 219493 [details, diff] grisbi.ebuild.diff Diff for reviewing changes easier
CCing Remi since he was interested on grisbi ( http://bugs.gentoo.org/show_bug.cgi?id=207605#c19 ), I am going to update also Guillaume mail in metadata Remi, Guillaume, seems that Guillaume's ebuild (attached one without commented lines) works ok for installing grisbi, but there are two issues I have seen: 1. It segfaults when merged with USE="cunit" -> Would be nice if you could send this problem upstream, even if this can be workarounded for now simply disabling cuna support... 2. This is really a "cosmetic" problem but, menu entry has no icon -> I think that it's because of having "Icon=grisbi.png" instead of "Icon=grisbi" in its desktop file. Thanks
For cunit : reported upstream. http://grisbi.tuxfamily.org/mantis/view.php?id=954 Answer is clear, cunit tests was a try and the test suite is currently not maintained, this is why the it fails. So it should be disabled and cunit USE flag removed.
Ok, I've committed an ebuild to portage which is Guillaume's with a couple improvements, among which : - drop USE=cunit - EAPI=2, which allowed me to drop a couple lines of code - "emake install" instead of "einstall" - patch to make it build with ~arch gcc - only grep/sed Makefile.am since we now run eautoreconf I've also started cleaning up grisbi in a local git repo, but that's not in portage. (and it will probably never be unless it gets accepted upstream) Cheers
Thanks a lot Remi
