Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Tom de Vries <tdevries@suse.de>
To: Sergio Durigan Junior <sergiodj@redhat.com>
Cc: gdb-patches@sourceware.org
Subject: Re: buildbot config cc-with-index not functional
Date: Wed, 22 May 2019 09:49:00 -0000	[thread overview]
Message-ID: <88f2db5f-7130-782c-c679-329b4a8e12fb@suse.de> (raw)
In-Reply-To: <871s0ru17t.fsf@redhat.com>

On 21-05-19 20:11, Sergio Durigan Junior wrote:
> On Monday, May 13 2019, Tom de Vries wrote:
> 
>> On 03-05-19 22:50, Sergio Durigan Junior wrote:
>>> On Friday, May 03 2019, Tom de Vries wrote:
>>>
>>>> On 03-05-19 21:47, Sergio Durigan Junior wrote:
>>>>> On Friday, May 03 2019, Tom de Vries wrote:
>>>>>
>>>>>> [ was: Re: [committed][gdb/testsuite] Add cc-with-gdb-index.exp ]
>>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> I've just added a target board cc-with-gdb-index.exp, to make it easy to
>>>>>> test using cc-with-tweaks.sh -i.
>>>>>>
>>>>>> Then I stumbled upon the build bot config Fedora-x86_64-cc-with-index,
>>>>>> and I took a look at the test results.
>>>>>>
>>>>>> I noticed that a test failing with cc-with-gdb-index.exp,
>>>>>> gdb.base/cache-index.exp (see PR24516 - "cc-with-gdb-index failures") is
>>>>>> passing in the cc-with-index config.
>>>>>>
>>>>>> I managed to reproduce that pass using the way the config uses
>>>>>> cc-with-tweaks.sh:
>>>>>> ...
>>>>>> $ make -k check \
>>>>>>   'CC_FOR_TARGET=/bin/sh gdb/contrib/cc-with-tweaks.sh -i gcc' \
>>>>>>   'CXX_FOR_TARGET=/bin/sh gdb/contrib/cc-with-tweaks.sh -i g++' \
>>>>>>   -j4 \
>>>>>>   FORCE_PARALLEL=1 \
>>>>>>   TS=1
>>>>>> ...
>>>>>>
>>>>>> So, AFAIU, the cc-with-index buildbot config does not actually test
>>>>>> using cc-with-tweaks.sh -i, as it intends to.
>>>>>
>>>>> Ah...  Thanks for bringing this up.  IIUC, CC_FOR_TARGET and
>>>>> CXX_FOR_TARGET need to be passed inside RUNTESTFLAGS.  I think that's
>>>>> the problem.  Can you please check if you can reproduce the bug when you
>>>>> do that?
>>>>>
>>>>
>>>> Confirmed, using:
>>>> ...
>>>> $ make -k check "RUNTESTFLAGS='CC_FOR_TARGET=/bin/sh
>>>> $pwd/gdb/contrib/cc-with-tweaks.sh -i gcc' 'CXX_FOR_TARGET=/bin/sh
>>>> $pwd/gdb/contrib/cc-with-tweaks.sh -i g++' gdb.base/index-cache.exp"
>>>> ...
>>>> I can reproduce the bug.
>>>
>>> Thanks.
>>>
>>>>> Also, I'm happy to know there's now a cc-with-gdb-index.exp flag.  So the
>>>>> buildbot can just use RUNTESTFLAGS='--target_board
>>>>> cc-with-gdb-index.exp', right?
>>>>>
>>>>
>>>> Yep, that's the idea.
>>>
>>> Thanks, I'll update the config file now.
>>>
>>
>> Hi,
>>
>> things still seem to be failing (
>> https://gdb-build.sergiodj.net/builders/Fedora-x86_64-cc-with-index/builds/12689/steps/test%20gdb/logs/stdio
>> ):
>> ...
>> make -k check 'RUNTESTFLAGS=CC_FOR_TARGET="/bin/sh
>> /home/gdb-buildbot/fedora-x86-64-3/fedora-x86-64-cc-with-index/binutils-gdb/gdb/contrib/cc-with-tweaks.sh
>> -i gcc" CXX_FOR_TARGET="/bin/sh
>> /home/gdb-buildbot/fedora-x86-64-3/fedora-x86-64-cc-with-index/binutils-gdb/gdb/contrib/cc-with-tweaks.sh
>> -i g++"' -j4 FORCE_PARALLEL=1 TS=1
>>    ...
>> make[1]: *** No rule to make target 'gcc CXX_FOR_TARGET=/bin/sh'.
>> make[1]: *** No rule to make target 'g++'.
>> ...
>> ?
> 
> Sorry, it's taken longer than I expected to get back to this, and then I
> had to struggle with quoting and argument-passing.

[ FWIW, I think it would have been easier to use the cc-with-gdb-index
target board. ]

> Anyway, I think I
> figured it out now.  Here's an example of the log file generated:
> 
>   https://gdb-build.sergiodj.net/results/Fedora-x86_64-cc-with-index/e6/e6f65e7573a317ac4efff26fe0e49fe1b9e7a596/gdb.log.xz
> 
> I'm still seeing a lot of "ERROR" messages, though:
> 
>   https://gdb-build.sergiodj.net/builders/Fedora-x86_64-cc-with-index/builds/12827/steps/test%20gdb/logs/stdio
> 
> I don't know if that's something related to the buildslave, or to
> something else.
> 

I cannot make too much sense out of it, with the parallel check weaving
output. Perhaps it would be easier to understand these ERRORs with make -O.

> Anyway, can you double check to see if I didn't forget something,
> please?

A test that is passing in normal testing, but is failing with
cc-with-gdb-index target board, is gdb.dwarf2/main-subprogram.exp (PR24549).

And I found that error in the logs, so I think that confirms that you
got it working.

Thanks,
- Tom


      reply	other threads:[~2019-05-22  9:49 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-05-03 10:58 [committed][gdb/testsuite] Add cc-with-gdb-index.exp Tom de Vries
2019-05-03 18:11 ` buildbot config cc-with-index not functional Tom de Vries
2019-05-03 19:47   ` Sergio Durigan Junior
2019-05-03 20:26     ` Tom de Vries
2019-05-03 20:50       ` Sergio Durigan Junior
2019-05-13 10:07         ` Tom de Vries
2019-05-21 18:11           ` Sergio Durigan Junior
2019-05-22  9:49             ` Tom de Vries [this message]

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=88f2db5f-7130-782c-c679-329b4a8e12fb@suse.de \
    --to=tdevries@suse.de \
    --cc=gdb-patches@sourceware.org \
    --cc=sergiodj@redhat.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