* [Qemu-devel] [PATCH] configure: disallow ccache during compile tests
@ 2015-10-28 17:56 John Snow
2015-10-29 6:17 ` Markus Armbruster
2015-11-02 14:57 ` Paolo Bonzini
0 siblings, 2 replies; 6+ messages in thread
From: John Snow @ 2015-10-28 17:56 UTC (permalink / raw)
To: qemu-devel; +Cc: peter.maydell, John Snow, pbonzini
If the user is using CCACHE during the configuration step,
it may interfere with some of the configuration tests,
particularly the "Is CCACHE interfering with macro analysis" step,
which is a bit of a poetic problem.
1) Disallow CCACHE from reading from the cache during configure,
but don't disable it to allow us to see if it causes other problems.
2) Force off CCACHE_CPP2 during the ccache test to get a deterministic
answer over whether or not we need to enable that feature later.
Signed-off-by: John Snow <jsnow@redhat.com>
---
configure | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/configure b/configure
index 7a1d08d..9c726eb 100755
--- a/configure
+++ b/configure
@@ -8,6 +8,9 @@
CLICOLOR_FORCE= GREP_OPTIONS=
unset CLICOLOR_FORCE GREP_OPTIONS
+# Don't allow CCACHE, if present, to use cached results of compile tests!
+export CCACHE_RECACHE=yes
+
# Temporary directory used for files created while
# configure runs. Since it is in the build directory
# we can safely blow away any previous version of it
@@ -4412,6 +4415,7 @@ fi
# check if ccache is interfering with
# semantic analysis of macros
+unset CCACHE_CPP2
ccache_cpp2=no
cat > $TMPC << EOF
static const int Z = 1;
--
2.4.3
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: disallow ccache during compile tests
2015-10-28 17:56 [Qemu-devel] [PATCH] configure: disallow ccache during compile tests John Snow
@ 2015-10-29 6:17 ` Markus Armbruster
2015-10-29 15:53 ` John Snow
2015-11-02 14:57 ` Paolo Bonzini
1 sibling, 1 reply; 6+ messages in thread
From: Markus Armbruster @ 2015-10-29 6:17 UTC (permalink / raw)
To: John Snow; +Cc: peter.maydell, qemu-devel, pbonzini
John Snow <jsnow@redhat.com> writes:
> If the user is using CCACHE during the configuration step,
> it may interfere with some of the configuration tests,
> particularly the "Is CCACHE interfering with macro analysis" step,
> which is a bit of a poetic problem.
>
> 1) Disallow CCACHE from reading from the cache during configure,
> but don't disable it to allow us to see if it causes other problems.
This is confusing. "don't disable it entirely"?
> 2) Force off CCACHE_CPP2 during the ccache test to get a deterministic
> answer over whether or not we need to enable that feature later.
>
> Signed-off-by: John Snow <jsnow@redhat.com>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: disallow ccache during compile tests
2015-10-29 6:17 ` Markus Armbruster
@ 2015-10-29 15:53 ` John Snow
2015-10-29 16:22 ` Markus Armbruster
0 siblings, 1 reply; 6+ messages in thread
From: John Snow @ 2015-10-29 15:53 UTC (permalink / raw)
To: Markus Armbruster; +Cc: peter.maydell, qemu-devel, pbonzini
On 10/29/2015 02:17 AM, Markus Armbruster wrote:
> John Snow <jsnow@redhat.com> writes:
>
>> If the user is using CCACHE during the configuration step,
>> it may interfere with some of the configuration tests,
>> particularly the "Is CCACHE interfering with macro analysis" step,
>> which is a bit of a poetic problem.
>>
>> 1) Disallow CCACHE from reading from the cache during configure,
>> but don't disable it to allow us to see if it causes other problems.
>
> This is confusing. "don't disable it entirely"?
>
We allow it to submit items into the cache, but not to read items from
the cache. This prevents it from picking up cached results to the
compile tests.
I don't disable it entirely (just pass-through to the real compiler
without using ccache at all) because I still want to allow ccache to be
processing things to test for ccache failures -- particularly the ccache
macro check that occurs later.
If I just disable ccache, that later check is meaningless. Setting the
cache to "write only" gives me the best of both worlds.
>> 2) Force off CCACHE_CPP2 during the ccache test to get a deterministic
>> answer over whether or not we need to enable that feature later.
>>
>> Signed-off-by: John Snow <jsnow@redhat.com>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: disallow ccache during compile tests
2015-10-29 15:53 ` John Snow
@ 2015-10-29 16:22 ` Markus Armbruster
2015-11-02 14:54 ` Paolo Bonzini
0 siblings, 1 reply; 6+ messages in thread
From: Markus Armbruster @ 2015-10-29 16:22 UTC (permalink / raw)
To: John Snow; +Cc: peter.maydell, qemu-devel, pbonzini
John Snow <jsnow@redhat.com> writes:
> On 10/29/2015 02:17 AM, Markus Armbruster wrote:
>> John Snow <jsnow@redhat.com> writes:
>>
>>> If the user is using CCACHE during the configuration step,
>>> it may interfere with some of the configuration tests,
>>> particularly the "Is CCACHE interfering with macro analysis" step,
>>> which is a bit of a poetic problem.
>>>
>>> 1) Disallow CCACHE from reading from the cache during configure,
>>> but don't disable it to allow us to see if it causes other problems.
>>
>> This is confusing. "don't disable it entirely"?
>>
>
> We allow it to submit items into the cache, but not to read items from
> the cache. This prevents it from picking up cached results to the
> compile tests.
>
> I don't disable it entirely (just pass-through to the real compiler
> without using ccache at all) because I still want to allow ccache to be
> processing things to test for ccache failures -- particularly the ccache
> macro check that occurs later.
>
> If I just disable ccache, that later check is meaningless. Setting the
> cache to "write only" gives me the best of both worlds.
All right, shows that I can confuse with the best of them!
It took my a moment to understand what "disallow but don't disable"
means. I think inserting "entirely" into your sentence right after
"disable it" makes it easier to understand.
>>> 2) Force off CCACHE_CPP2 during the ccache test to get a deterministic
>>> answer over whether or not we need to enable that feature later.
>>>
>>> Signed-off-by: John Snow <jsnow@redhat.com>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: disallow ccache during compile tests
2015-10-29 16:22 ` Markus Armbruster
@ 2015-11-02 14:54 ` Paolo Bonzini
0 siblings, 0 replies; 6+ messages in thread
From: Paolo Bonzini @ 2015-11-02 14:54 UTC (permalink / raw)
To: Markus Armbruster, John Snow; +Cc: peter.maydell, qemu-devel
On 29/10/2015 17:22, Markus Armbruster wrote:
> John Snow <jsnow@redhat.com> writes:
>
>> On 10/29/2015 02:17 AM, Markus Armbruster wrote:
>>> John Snow <jsnow@redhat.com> writes:
>>>
>>>> If the user is using CCACHE during the configuration step,
>>>> it may interfere with some of the configuration tests,
>>>> particularly the "Is CCACHE interfering with macro analysis" step,
>>>> which is a bit of a poetic problem.
>>>>
>>>> 1) Disallow CCACHE from reading from the cache during configure,
>>>> but don't disable it to allow us to see if it causes other problems.
>>>
>>> This is confusing. "don't disable it entirely"?
>>>
>>
>> We allow it to submit items into the cache, but not to read items from
>> the cache. This prevents it from picking up cached results to the
>> compile tests.
>>
>> I don't disable it entirely (just pass-through to the real compiler
>> without using ccache at all) because I still want to allow ccache to be
>> processing things to test for ccache failures -- particularly the ccache
>> macro check that occurs later.
>>
>> If I just disable ccache, that later check is meaningless. Setting the
>> cache to "write only" gives me the best of both worlds.
>
> All right, shows that I can confuse with the best of them!
>
> It took my a moment to understand what "disallow but don't disable"
> means. I think inserting "entirely" into your sentence right after
> "disable it" makes it easier to understand.
Ok, commit log adjusted.
Paolo
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Qemu-devel] [PATCH] configure: disallow ccache during compile tests
2015-10-28 17:56 [Qemu-devel] [PATCH] configure: disallow ccache during compile tests John Snow
2015-10-29 6:17 ` Markus Armbruster
@ 2015-11-02 14:57 ` Paolo Bonzini
1 sibling, 0 replies; 6+ messages in thread
From: Paolo Bonzini @ 2015-11-02 14:57 UTC (permalink / raw)
To: John Snow, qemu-devel; +Cc: peter.maydell
On 28/10/2015 18:56, John Snow wrote:
> If the user is using CCACHE during the configuration step,
> it may interfere with some of the configuration tests,
> particularly the "Is CCACHE interfering with macro analysis" step,
> which is a bit of a poetic problem.
>
> 1) Disallow CCACHE from reading from the cache during configure,
> but don't disable it to allow us to see if it causes other problems.
>
> 2) Force off CCACHE_CPP2 during the ccache test to get a deterministic
> answer over whether or not we need to enable that feature later.
>
> Signed-off-by: John Snow <jsnow@redhat.com>
> ---
> configure | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/configure b/configure
> index 7a1d08d..9c726eb 100755
> --- a/configure
> +++ b/configure
> @@ -8,6 +8,9 @@
> CLICOLOR_FORCE= GREP_OPTIONS=
> unset CLICOLOR_FORCE GREP_OPTIONS
>
> +# Don't allow CCACHE, if present, to use cached results of compile tests!
> +export CCACHE_RECACHE=yes
> +
> # Temporary directory used for files created while
> # configure runs. Since it is in the build directory
> # we can safely blow away any previous version of it
> @@ -4412,6 +4415,7 @@ fi
> # check if ccache is interfering with
> # semantic analysis of macros
>
> +unset CCACHE_CPP2
> ccache_cpp2=no
> cat > $TMPC << EOF
> static const int Z = 1;
>
Queued, thanks.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2015-11-02 14:57 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-10-28 17:56 [Qemu-devel] [PATCH] configure: disallow ccache during compile tests John Snow
2015-10-29 6:17 ` Markus Armbruster
2015-10-29 15:53 ` John Snow
2015-10-29 16:22 ` Markus Armbruster
2015-11-02 14:54 ` Paolo Bonzini
2015-11-02 14:57 ` Paolo Bonzini
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).