* Testsuite failing
@ 2014-07-09 4:24 Matthieu Vachon
2014-07-09 4:47 ` Matthieu Vachon
2014-07-09 22:03 ` Per Bothner
0 siblings, 2 replies; 6+ messages in thread
From: Matthieu Vachon @ 2014-07-09 4:24 UTC (permalink / raw)
To: kawa
[-- Attachment #1: Type: text/plain, Size: 246 bytes --]
Here a patch that make `testsuite/errors3.scm` pass. The message was
expecting `... with required type int` while it's now `... with
required type character`.
Maybe the code should be fixed and not the test ... not sure about it.
Regards,
Matt
[-- Attachment #2: fix-testsuite-errors3.patch --]
[-- Type: application/octet-stream, Size: 645 bytes --]
Index: testsuite/errors3.scm
===================================================================
--- testsuite/errors3.scm (revision 7955)
+++ testsuite/errors3.scm (working copy)
@@ -4,5 +4,5 @@
(format #t "Missing exception i: ~w~%" i))
(ex java.lang.Exception
(format #t "Caught: ~w~%" ex)))
-;; Diagnostic: errors3.scm:2:1: warning - type boolean is incompatible with required type int
+;; Diagnostic: errors3.scm:2:1: warning - type boolean is incompatible with required type character
;; Output: Caught: Value 'true' for variable 'i' has wrong type (java.lang.Boolean) (java.lang.Boolean cannot be cast to java.lang.Number)
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: Testsuite failing
2014-07-09 4:24 Testsuite failing Matthieu Vachon
@ 2014-07-09 4:47 ` Matthieu Vachon
2014-07-09 4:49 ` Matthieu Vachon
2014-07-09 22:03 ` Per Bothner
1 sibling, 1 reply; 6+ messages in thread
From: Matthieu Vachon @ 2014-07-09 4:47 UTC (permalink / raw)
To: kawa
The testsuite fails on Java 7 with the patch not applied. But with the
patch applied, it instead fails on Java 8
(https://travis-ci.org/maoueh/kawa-fork/jobs/29473740).
It might be a problem with OpenJDK8 though. Testing using oracle jdk8
would need to be performed to be sure.
Conclusion, patch is not enough/good.
Matt
On Wed, Jul 9, 2014 at 12:24 AM, Matthieu Vachon
<matthieu.o.vachon@gmail.com> wrote:
> Here a patch that make `testsuite/errors3.scm` pass. The message was
> expecting `... with required type int` while it's now `... with
> required type character`.
>
> Maybe the code should be fixed and not the test ... not sure about it.
>
> Regards,
> Matt
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Testsuite failing
2014-07-09 4:24 Testsuite failing Matthieu Vachon
2014-07-09 4:47 ` Matthieu Vachon
@ 2014-07-09 22:03 ` Per Bothner
2014-07-10 1:34 ` Per Bothner
1 sibling, 1 reply; 6+ messages in thread
From: Per Bothner @ 2014-07-09 22:03 UTC (permalink / raw)
To: kawa
On 07/08/2014 09:24 PM, Matthieu Vachon wrote:
> Here a patch that make `testsuite/errors3.scm` pass. The message was
> expecting `... with required type int` while it's now `... with
> required type character`.
>
> Maybe the code should be fixed and not the test ... not sure about it.
The test us fine. The problem is the code in kawa.standard.Scheme#formatTest.
Specifically how it enters both T and T.getImplementationType() into the
typeToStringMap. Thia looses if T is the character type, since the
implementation type for T is int.
--
--Per Bothner
per@bothner.com http://per.bothner.com/
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Testsuite failing
2014-07-09 22:03 ` Per Bothner
@ 2014-07-10 1:34 ` Per Bothner
2014-07-10 2:09 ` Matthieu Vachon
0 siblings, 1 reply; 6+ messages in thread
From: Per Bothner @ 2014-07-10 1:34 UTC (permalink / raw)
To: kawa
On 07/09/2014 03:02 PM, Per Bothner wrote:
>
>
> On 07/08/2014 09:24 PM, Matthieu Vachon wrote:
>> Here a patch that make `testsuite/errors3.scm` pass. The message was
>> expecting `... with required type int` while it's now `... with
>> required type character`.
>>
>> Maybe the code should be fixed and not the test ... not sure about it.
>
> The test us fine. The problem is the code in kawa.standard.Scheme#formatTest.
> Specifically how it enters both T and T.getImplementationType() into the
> typeToStringMap. Thia looses if T is the character type, since the
> implementation type for T is int.
I checked in a fix.
(If it wasn't clear: The reason for the Java 7 vs Java 8 inconsistency
isn't anything "real". Rather it is due to unspecified traversal order
of hash tables. Somehow Java 8 tweaked the implementation enough so
the bug was hidden on Java 8.)
--
--Per Bothner
per@bothner.com http://per.bothner.com/
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: Testsuite failing
2014-07-10 1:34 ` Per Bothner
@ 2014-07-10 2:09 ` Matthieu Vachon
0 siblings, 0 replies; 6+ messages in thread
From: Matthieu Vachon @ 2014-07-10 2:09 UTC (permalink / raw)
To: Per Bothner; +Cc: kawa
On Wed, Jul 9, 2014 at 9:34 PM, Per Bothner <per@bothner.com> wrote:
>
>
> I checked in a fix.
>
> (If it wasn't clear: The reason for the Java 7 vs Java 8 inconsistency
> isn't anything "real". Rather it is due to unspecified traversal order
> of hash tables. Somehow Java 8 tweaked the implementation enough so
> the bug was hidden on Java 8.)
>
Thanks for the follow-up, I was indeed wondering why it was working
depending on the platform.
Matt
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2014-07-10 2:09 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-07-09 4:24 Testsuite failing Matthieu Vachon
2014-07-09 4:47 ` Matthieu Vachon
2014-07-09 4:49 ` Matthieu Vachon
2014-07-09 22:03 ` Per Bothner
2014-07-10 1:34 ` Per Bothner
2014-07-10 2:09 ` Matthieu Vachon
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 read-only IMAP folder(s) and NNTP newsgroup(s).