Summary: | app-office/openoffice-3.0.0 and -3.1.0_beta6 fail to build with either java or opengl useflag | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Thomas Sachau <tommy> |
Component: | New packages | Assignee: | Gentoo Office Team <office> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | spatz |
Priority: | High | ||
Version: | 2008.0 | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Thomas Sachau
![]() I tried building version 3.1.0 on amd64 and got the same segfault. I'm using --as-needed as well, so I'm thinking that's the problem. I'll try rebuilding without it and see how that goes. (In reply to comment #1) > I tried building version 3.1.0 on amd64 and got the same segfault. I'm using > --as-needed as well, so I'm thinking that's the problem. I'll try rebuilding > without it and see how that goes. > 3.1.0? Where did you get that? :-) I tried without setting any LDFLAGS, but the result was the same for me. With empty LDFLAGS I don't get a segfault, but the build still fails because lightproof-1.0.oxt doesn't exist, whatever that means :/ It could be that I forgot to change something in the ebuild when I bumped to 3.1.0, but I did change MY_PV. Did you try this with a recent release of OOo 3.1.0? (In reply to comment #4) > Did you try this with a recent release of OOo 3.1.0? > Yes, still parallel build issue (so currenlty forced to use -j1) and still the segfault with either opengl or java. I did try to run that checkdll binary with the so as argument in gdb, but the backtrace is not complete, so i fear that this binary is not compiled, but part of the source tarball. Since my time is a bit limited, that is all i found out until now. But if you have special testcases or other questions, i can surely run another ooo compile over night. Version 3.1.0 with both java and opengl enabled compiles and runs fine for me on ~amd64. Is this bug still relevant? (In reply to comment #6) > Version 3.1.0 with both java and opengl enabled compiles and runs fine for me > on ~amd64. Is this bug still relevant? > Still happens to me with 3.1.0.6 Still getting this with 3.1.1? Guess this is hardened related, never got it with non-hardened system... (In reply to comment #8) > Still getting this with 3.1.1? Guess this is hardened related, never got it > with non-hardened system... > This issue is related to hardened, it seems like this checkdll binary does some in-memory actions, which are prevented by PaX. lo-3.4.2.2 seems to run on hardened. Please reopen this bug if you still have compilation failures. Thanks |