Re: bash: how to interpret the test-results

On 2/14/25 2:48 PM, Rainer Fiebig ([email protected] via lfs-support Mailing List) wrote:
> I must admit that I've never paid attention to the test-results for
> bash.  Bash has always worked, no problems.
> 
> After updating glibc to -2.41, I rebuilt bash-5.1.16 and ran the tests
> with  "make tests"  (as described in "Install" in the sources).
> 
> The book says: "Any output from diff (prefixed with < and >) indicates a
> test failure, unless there is a message saying the difference can be
> ignored."
> 
> A line in the log says: "Any output from any test, unless otherwise
> noted, indicates a possible anomaly"
> 
> According to that almost all of the 82 tests ("run-soandso") must have
> failed because all have "anomalies".  For example:
> 
> run-alias
> 3c3
> < ./alias.tests: Zeile 38: qfoo: Kommando nicht gefunden.
> ---
>> ./alias.tests: line 38: qfoo: command not found
> run-appendop
> 18c18
> < ./appendop.tests: Zeile 97: x: Schreibgeschützte Variable.
> ---
>> ./appendop.tests: line 97: x: readonly variable
> run-arith
> 63c63
> < ./arith.tests: Zeile 127: 1 ? 20 : x+=2: Versuchte Zuweisung zu etwas,
> das keine Variable ist. (Fehlerverursachendes Zeichen ist "+=2").
> ---
>> ./arith.tests: line 127: 1 ? 20 : x+=2: attempted assignment to
> non-variable (error token is "+=2")
> 83,92c83,92
> 
> And so on and so on.  "<" in German, ">" in English.  The test-results
> are the same with glibc-2.35 and -2.41.  I mean I had _zero_ FAILS of
> over 6000 tests for glibc-2.41 but almost every test for bash indicates
> failure?
> 
> I don't know what to make of it.  Any insights, please?  Thanks.

It sounds like a language issue to me.  Did you try setting LANG=C for the tests?

   -- Bruce

-- 
http://lists.linuxfromscratch.org/sympa/info/lfs-support
Unsubscribe: See the above information page