I have now released version 0.3 of Qalculate! and I think that it deserves an ebuild. Qalculate! is a powerful multi-purpose desktop calculator for GNU/Linux. For more information visit http://qalculate.sourceforge.net. Reproducible: Always Steps to Reproduce:
Created attachment 19601 [details] qalculate-gtk-0.3.ebuild
The screenshots look nice... suggestion for the ebuild: use RDEPEND="something" and DEPEND="${RDEPEND} something_more" to avoid redundancies.
Created attachment 19855 [details] qalculate-gtk-0.3.1.ebuild New version, updated dependencies and support for nls use flag.
Created attachment 24923 [details] Version bump This as compiled fine on my system. There is no minimal version requierment other than gtp and glade, so why provoque uneccessary updates? :)
Created attachment 26200 [details] qalculate-gtk-0.4.1.ebuild - new version - remove version requirement for wget - lower version requirement for gnuplot (3.7 is definitely needed) - some cosmetic changes - cln do need to be >= 1.1 (I'm the main developer so I should know)
Needs XML-Parser as well...
not exactly... it needs intltool which depends on XML-Parser.
oops. my mistake. intltool is not needed for compilation.
Created attachment 26963 [details] qalculate-gtk-0.4.2a.ebuild add XML-Parser dependency and, while I'm at it, update to the upcoming new version which needs scrollkeeper and optionally libgnome.
Created attachment 28431 [details] qalculate-gtk-0.5.ebuild Well, yours seem a bit more elaborate than mine. I'm too lazy to synced them. The latest stable gnuplot available in the portage tree seem to work fine here (it's 3.something).
Created attachment 28433 [details] qalculate-gtk-0.6.1.ebuild A more complete version.
I tested qalculate-gtk-0.5.1.ebuild on my system and it runs fine. Nice calc.
I updated to 0.6 (and I give up on my ebuild version -- it's not every ebuilds that are maintained by the author of the sw, so why get in the way ;) ). Suggestion, if I may: Maybe you should either link this bug to where reside your ebuild tarball, or link your sourceforge page to this bug and upload your tarball here. The real question is, howcome this fine software did not make it to the main portage tree?
Mostly the reason is simple: Lack of time.
qalculate 0.6.1 segfaults at startup. btw, doesn't it have a dependency on giac ?
no. it does not have any dependency on giac (v. 0.5.1 could optionally use giac for some functions such as integration, but that has not yet been enabled in the 0.6 series). Regarding the segfault, do you get any error messages?, do you use any aggressive CXXFLAGS/CFLAGS? (does it help compiling without them?), do you have any old .qalculate directory? (does it help removing it), do you use any fancy gcc version? ...
no error message. i've rebuilded qalculate without "-Os". now qalculate starts and seems to run fine, however it segfaults when i close it. you should filter out the -Os flag.
Created attachment 35110 [details] qalculate-gtk-0.6.2.ebuild replace -Os with -O2
regarding segfault on close: on exit Qalculate saves mode and definitions, and terminates running threads. so if you could try each of these actions by saving mode and definitions from the File and Mode menus, and type in for example "googleplex" and then press Stop in the progress dialog that pops up. (this is probably not related to the ebuild and should be moved to another forum if it goes on)
the segfault on exit has been fixed in cvs
thanks.
renamed ebuild for 0.6.2: it builds fine. no segfault when closing. everything seems to be good, thanks! commiting as ~x86 would be great.
Added to ~x86 with a small change: I added "inherit flag-o-matic" since the replace-flags function is used. Lots of improvements since I last tested this piece of software!
