[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Bug ld/23727] ld with custom sysroot ignores original rpath
From: |
nehaljw.kkd1 at gmail dot com |
Subject: |
[Bug ld/23727] ld with custom sysroot ignores original rpath |
Date: |
Mon, 01 Oct 2018 21:57:28 +0000 |
https://sourceware.org/bugzilla/show_bug.cgi?id=23727
Nehal J Wani <nehaljw.kkd1 at gmail dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |UNCONFIRMED
Resolution|INVALID |---
--- Comment #3 from Nehal J Wani <nehaljw.kkd1 at gmail dot com> ---
I read through the conversation on the mailing list.
Further observations:
- If I do add -Wl,-rpath-link,$PWD/libs the link succeeds.
- If I use the gold linker, the link succeeds. -Wl,-fuse-ld=gold
If you think that not searching out of the sysroot at all is the correct
behavior, then why is it restricted only to -rpath and not applied to
-rpath-link? Why is the behavior of the cross gold linker different?
My use case is simple. I want to cross compile let's say libssh2 for a target
architecture and I have the libcrypto.so and libz.so dependencies already
compiled for the target architecture kept in /path/to/some/prefix/libs. While
compiling libssh2, I want to link against those libraries. Why am I expected to
keep them in the sysroot always? Does it not defeat the whole purpose of having
-rpath in the first place? Does this not seem like a valid use case?
--
You are receiving this mail because:
You are on the CC list for the bug.