Re: Test failures

From: Vladimir Prus (ghost_at_[hidden])
Date: 2003-01-05 13:49:27

David Abrahams wrote:
>>I see "a.o" here. Did you look at preserved test directory and found no
>>"a.o" there?
> 
> 
> I didn't know about the PRESERVE hack. Where can I find out about
> these.

In "test_system.html", under "Test result" section.

> No, I modified BoostBuild.py to do an ls -R at the point of the error,
> which obviously didn't work. When I use PRESERVE, there it is:
> 
> y.x2
> y.x1
> y.o
> y.cpp
> x.o
> x.c
> main-target-obj_2
> main-target-obj_1
> f.exe
> e.target.cpp
> e.positions
> e.o
> e.marked.cpp
> e.exe
> d_parser.whl
> d_parser.o
> d_parser.lr0
> d_parser.cpp
> d_lexer.o
> d_lexer.dlp
> d_lexer.cpp
> c.o
> c.h
> c.cpp
> b.o
> a.o <=======
> a.exe
> 
> I have no clue why it's reported as missing.

Can you dive in "generators-test" directory and see if there's
"bin" directory there. If "a.o" existed before the test run,
then if would be reported as "not added". It would be better
to differentiate between "was present before" and "was not present before",
just I did not code that yet.

If that fails, try calling pprint method of "tree" and "previous_tree"
attributes of the Tester class, somewhere around the end of "run_build_system".

- Volodya

- Volodya