Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Tom de Vries via Gdb-patches <gdb-patches@sourceware.org>
To: Joel Brobecker <brobecker@adacore.com>
Cc: Tom Tromey <tom@tromey.com>, gdb-patches@sourceware.org
Subject: Re: [PATCH][gdb/testsuite] Handle unrecognized command line option in gdb_compile_test
Date: Wed, 8 Sep 2021 15:45:40 +0200	[thread overview]
Message-ID: <cb31b1f3-7f94-6de0-0a5b-4c7ffb2b5270@suse.de> (raw)
In-Reply-To: <20210908122232.GE9184@adacore.com>

On 9/8/21 2:22 PM, Joel Brobecker wrote:
>> When running the gdb testsuite with gnatmake-4.8, I get many fails of the
>> following form:
>> ...
>> gcc: error: unrecognized command line option '-fgnat-encodings=all'^M
>> gnatmake: "gdb.ada/O2_float_param/foo.adb" compilation error^M
>> compiler exited with status 1
>> compilation failed: gcc ... gdb.ada/O2_float_param/foo.adb
>> gcc: error: unrecognized command line option '-fgnat-encodings=all'
>> gnatmake: "gdb.ada/O2_float_param/foo.adb" compilation error
>> FAIL: gdb.ada/O2_float_param.exp: scenario=all: compilation foo.adb
>> ...
>>
>> Fix this by marking the test unsupported instead, such that we have:
>> ...
>> UNSUPPORTED: gdb.ada/O2_float_param.exp: scenario=all: compilation foo.adb \
>>   (unsupported option '-fgnat-encodings=all')
>> ...
>>
>> Also stop gdb_compile_test from doing pass/fail.  The gdb testsuite is meant
>> to test gdb, and a fail is meant to indicate a problem with gdb.  A compiler
>> failure means that the test is unsupported.
> 
> For the last part, I suggest we remain consistent with what we do when
> building programs in other languages. Looking at gdb_compile_test,
> we have ...
> 
>     } else {
>         verbose -log "compilation failed: $output" 2
>         fail "compilation [file tail $src]"
> 

Right, that's what it says in gdb_compile_test. Using that proc however,
is the exception, used only for ada and f77.

Becoming consistent with other languages means to change that fail in
gdb_compile_test to unsupported.   Which is what this patch does.

> In my opinion, it's better to generate a FAIL here, because a failure
> to compile could come from an error in the testcase, rather than
> an unsupported build.
> 

[ To reiterate my earlier point, it's counter-pattern to generate a FAIL
for something that's not a problem in the tool-under-test.  So much
counter-pattern that I had to resist the reflex to commit this as obvious. ]

Indeed, there could be an error in the testcase.  Also, there could be
no error in the test-case.

In the former case, having a FAIL instead of UNSUPPORTED could increase
the change of noticing this, agreed.  In the latter case, having a FAIL
instead of UNSUPPORTED when testing an older compiler, is an annoyance
which drowns out FAILs that do need attention.

So, I'm not saying that noticing test-case errors causing compilation
problem is not worthwhile.  But I'm saying that turning the compilation
errors into FAILs is the wrong solution.

Thanks,
- Tom

>> Tested on x86_64-linux.
>>
>> Bug: https://sourceware.org/bugzilla/show_bug.cgi?id=28318
>>
>> Any comments?
>>
>> Thanks,
>> - Tom
>>
>> [gdb/testsuite] Handle unrecognized command line option in gdb_compile_test
>>
>> ---
>>  gdb/testsuite/lib/gdb.exp | 35 +++++++++++++++++++++++------------
>>  1 file changed, 23 insertions(+), 12 deletions(-)
>>
>> diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
>> index ff19760bac4..460b44c2160 100644
>> --- a/gdb/testsuite/lib/gdb.exp
>> +++ b/gdb/testsuite/lib/gdb.exp
>> @@ -2201,22 +2201,33 @@ proc gdb_interact { } {
>>  
>>  # Examine the output of compilation to determine whether compilation
>>  # failed or not.  If it failed determine whether it is due to missing
>> -# compiler or due to compiler error.  Report pass, fail or unsupported
>> -# as appropriate
>> +# compiler or due to compiler error.  Report unsupported as appropriate.
>>  
>>  proc gdb_compile_test {src output} {
>> +    set msg "compilation [file tail $src]"
>> +
>>      if { $output == "" } {
>> -	pass "compilation [file tail $src]"
>> -    } elseif { [regexp {^[a-zA-Z_0-9]+: Can't find [^ ]+\.$} $output] } {
>> -	unsupported "compilation [file tail $src]"
>> -    } elseif { [regexp {.*: command not found[\r|\n]*$} $output] } {
>> -	unsupported "compilation [file tail $src]"
>> -    } elseif { [regexp {.*: [^\r\n]*compiler not installed[^\r\n]*[\r|\n]*$} $output] } {
>> -	unsupported "compilation [file tail $src]"
>> -    } else {
>> -	verbose -log "compilation failed: $output" 2
>> -	fail "compilation [file tail $src]"
>> +	return
>> +    }
>> +
>> +    if { [regexp {^[a-zA-Z_0-9]+: Can't find [^ ]+\.$} $output]
>> +	 || [regexp {.*: command not found[\r|\n]*$} $output]
>> +	 || [regexp {.*: [^\r\n]*compiler not installed[^\r\n]*[\r|\n]*$} $output] } {
>> +	unsupported "$msg (missing compiler)"
>> +	return
>>      }
>> +
>> +    set gcc_re ".*: error: unrecognized command line option "
>> +    set clang_re ".*: error: unsupported option "
>> +    if { [regexp "(?:$gcc_re|$clang_re)(\[^ \t;\r\n\]*)" $output dummy option]
>> +	 && $option != "" } {
>> +	unsupported "$msg (unsupported option $option)"
>> +	return
>> +    }
>> +
>> +    # Unclassified compilation failure, be more verbose.
>> +    verbose -log "compilation failed: $output" 2
>> +    unsupported "$msg"
>>  }
>>  
>>  # Return a 1 for configurations for which we don't even want to try to
> 

  reply	other threads:[~2021-09-08 13:46 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-08 11:20 Tom de Vries via Gdb-patches
2021-09-08 12:22 ` Joel Brobecker
2021-09-08 13:45   ` Tom de Vries via Gdb-patches [this message]
2021-09-08 14:12     ` Joel Brobecker
2021-09-10 16:38 ` Tom de Vries via Gdb-patches

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=cb31b1f3-7f94-6de0-0a5b-4c7ffb2b5270@suse.de \
    --to=gdb-patches@sourceware.org \
    --cc=brobecker@adacore.com \
    --cc=tdevries@suse.de \
    --cc=tom@tromey.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox