This is the mail archive of the
mailing list for the glibc project.
In a cross-compile do host and build need exactly the same tools?
- From: Carlos O'Donell <carlos at redhat dot com>
- To: "Joseph S. Myers" <joseph at codesourcery dot com>, GNU C Library <libc-alpha at sourceware dot org>
- Date: Mon, 22 Oct 2018 11:35:58 -0400
- Subject: In a cross-compile do host and build need exactly the same tools?
This came up in the proposal discussion for python 3.4.
Given these lines in Rules:
293 # Static pattern rule that matches the test-* targets to their .c and .py
294 # prerequisites. It'll run the corresponding test script for each test program
295 # we compiled and place its output in the corresponding .out file.
296 # The pretty printer files and test_common_printers.py must be present for all.
297 $(tests-printers-out): $(objpfx)%.out: $(objpfx)% %.py %.c $(pretty-printers) \
299 $(test-wrapper-env) $(py-env) \
300 $(PYTHON) $*.py $*.c $(objpfx)$* $(pretty-printers) > $@; \
Particularly to execute the pretty printer tests written in python
wtih PExpect and gdb.
It looks like we would require python 3.4 on the host, even
in a cross-build environment to get clean test results.
Is it a general principle that we require exactly the same tools on build
and host to get clean test results?