This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[PATCH 0/3] GNAT test suite fixes for build sysroot


Hi,

 In the course of setting up GCC regression testing for the RISC-V target 
I have discovered that the GNAT test suite does not correctly respond to 
the test environment settings passed from the test harness in my setup and 
consequently no test case works correctly.

 In my particular setup `--with-build-sysroot=' has been used to configure 
GCC and also additional linker flags are defined in the target board so 
that adjusted paths are used for the dynamic loader and shared libraries 
to be found in the build sysroot at the test suite run time.  That plays 
well with most of the test suites included with GCC, but not the GNAT one.

 I have tracked down the cause to three issues spread across `gnatmake', 
the GNAT test driver and the DejaGNU target driver.  These issues are 
independent from each other, however all have to be addressed for the test 
suite to run, so I have decided to group them into a series despite that 
each of them can be applied separately, and the DejaGNU one goes to a 
separate project even.

 With them all in place I get these results with the `riscv64-linux-gnu'
target:

		=== gnat Summary ===

# of expected passes		2926
# of unexpected failures	5
# of expected failures		23
# of unsupported tests		26

which I think are pretty good, especially compared to the original results 
without these changes:

		=== gnat Summary ===

# of expected passes		1580
# of unexpected failures	133
# of unexpected successes	23
# of unresolved testcases	875
# of unsupported tests		26

 See individual change descriptions for details.

  Maciej

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]