The build for net-p2p/qtella-0.5.4 fails with gcc-3.2.1. full error attached.
Created attachment 6518 [details] Error message
Seems to be a dupe of 7285
The 0.5.3 builds also fail for me, but I don't think it's the new gcc, I think it's the KDE 3.1-rc5 header files that are the problem, as I've built these before successfully with the new gcc.
Yeah, i've been on this bug for quite some time now, send multiple mails to the author about it, which he doesn't seem to bother to answer me :o( I asked several programmers to look at it, but most get stuck.. I read about a guy who had a patch for this on the qtella-devel list, but his email address was partially hidden, so i could only see "younes@wa..." :o( It has something to do with #define KDE having to be #define HAVE_KDE and -DHAVE_KDE... The other error is due to an string conversion from QString to a normal string or something.. I'm stuck with this one as long as i don't receive a patch from somebody :'o( BTW, i have qtella 0.6.0 in my local cvs tree, but it has the same error..
Qtella 0.61 was released on 3 Feb 2003, and lists fixes for compiling against KDE 3.1 and Qt 3.1 in the change log.
*** Bug 15576 has been marked as a duplicate of this bug. ***
well, i copied qtella-0.5.4.ebuild to 0.6.1 and it doesn't compile. I get: QtellaSub.cpp: In constructor `QtellaSub::QtellaSub(QApplication*, const char*, unsigned int)': QtellaSub.cpp:799: `ui_checkbox_fileassociations' undeclared (first use this function) QtellaSub.cpp:799: (Each undeclared identifier is reported only once for each function it appears in.) QtellaSub.cpp:800: `ui_checkbox_systemtray' undeclared (first use this function) make[1]: *** [QtellaSub.o] Error 1 anybody has a fix for this?
I tried compiling from a .ebuild, and it compiles for me. Qt 3.1.0-r1, kdelibs 3.1. Unfortunately, the build doesn't actually run. Hmmm.
Created attachment 8491 [details] Qtella v 0.6.1 ebuild (net-p2p) This version compiles with newer qt (3.1) libs
Hannes, it works for me here (this is current kde cvs though), I'd commit if not for your comment... Tom, any more detailed diagnosis? Is it crashing with some signal? Try running strace on it perhaps.
I tried the ebuild, and it compiled successfully. It also seems to run fine, but only inside of KDE 3.1. If I try running the program from within fvwm, it hangs. Here's the tail end of an strace: brk(0x824c000) = 0x824c000 brk(0x824d000) = 0x824d000 brk(0x824e000) = 0x824e000 brk(0x824f000) = 0x824f000 brk(0x8250000) = 0x8250000 brk(0x8251000) = 0x8251000 brk(0x8252000) = 0x8252000 brk(0x8253000) = 0x8253000 fstat64(10, {st_mode=S_IFREG|0600, st_size=6282, ...}) = 0 munmap(0x40003000, 6282) = 0 rt_sigaction(SIGBUS, {SIG_DFL}, {0x4070b580, [BUS], SA_RESTART|0x4000000}, 8) = 0 close(10) = 0 brk(0x8254000) = 0x8254000 stat64("/home/tlcoll1/.kde3.1/share/config/Qtellarc", 0xbfffec90) = -1 ENOENT (No such file or directory) stat64("/usr/kde/3.1/share/config/Qtellarc", 0xbfffec90) = -1 ENOENT (No such file or directory) write(5, "\20\0\6\0\16\0\0\0KIPC_COMM_ATOM\0\0\20\0\7\0\22\0@\0"..., 80) = 80 read(5, "\1\0f\0\0\0\0\0\310\1\0\0\0\0\0\0\1\0\0\0\0\0\0\0\3703"..., 32) = 32 read(5, "\1\0g\0\0\0\0\0\311\1\0\0\0\0\0\0\0\0\0\0\0\0\0\0\3703"..., 32) = 32 read(5, "\1\0h\0\0\0\0\0\312\1\0\0\0\0\0\0\0\0\0\0\0\0\0\0\3703"..., 32) = 32 getpid() = 4921 uname({sys="Linux", node="alexis.smart.net", ...}) = 0 open("/home/tlcoll1/.DCOPserver_alexis.smart.net__0", O_RDONLY|O_LARGEFILE) = -1 ENOENT (No such file or directory) brk(0x8255000) = 0x8255000 access("/usr/kde/3.1/bin/kdeinit", F_OK) = 0 access("/usr/kde/3.1/bin/kdeinit", X_OK) = 0 lstat64("/usr/kde/3.1/bin/kdeinit", {st_mode=S_IFREG|0755, st_size=35456, ...}) = 0 write(5, "-\0\5\0\1\0 \1\6\0PCcursorAT^M\10\0\2\0 \1\1\0 \1"..., 68) = 68 fork() = 4924 wait4(4924, (Hangs in the wait4()). Not sure what's going on, but I'd say commit the ebuild, as I'm guessing this is something with qtella rather than with the ebuild.
Hmm... If compilnig without KDE support, Hannes' error shows up. I added a patch but it's just a workaround not a real fix. There is then a 'check' item in the qtella menu, which first takes up full cpu for a minute when clicked and then paints the text in all the qtella widgets in strange colors :-) I don't know what it's supposed to do. *shrug* BTW it seems the qtella (sourceforge) cvs isn't actively used between releases? There hasn't been a single commit there AFAICS since the 0.6.1 tarball. Anyway, I'm committing a 0.6.1 with this pathc to ~x86, since it at least works better than the current version (if you don't click that check thing). Will mail the qtella list as well.
They did not answer my email. Rather predictable. No changes on sf.net qtella cvs, no new release. The 'check' issue seems to be gone thought with qt 3.1.2 and kde 3.1.1a. Confirm please, I'd like to close these if possible.
the issue is still there, while qtella 0.6.1 is compiling good, maybe qtella 0.5.4 should be masked?
I marked qtella-0.6.1 stable on x86. please reopen if there are any problems left (I can start it in papuawm).