Hi Magnus,
herein I will be talking only about 1st two warnings.
although your analyze is correct, it doesn't take into account the fact that
the warnings report situations that can't happen in current codebase, and gcc
doesn't report them in our "regular" builds b/c it can proof that
- in mac
On 2018-08-31 01:20, Magnus Ihse Bursie wrote:
These are two different arguments for turning off warnings for code
coverage:
1) gcc is producing incorrect warnings
2) the warnings might be correct, but we are going to treat such bugs
as low priority
I understand and accept 1, but I do not
There are two aspects here that sound similar, but is not:
Erik says:
On Aug 30, 2018, at 6:26 AM, Erik Joelsson wrote:
I shared your opinion at first while discussing this offline with Leonid. What
changed my mind was the claim that the warnings cannot be truly trusted when
GCC is generatin
Hi
I read
It’d be much better and reliable to fix makefiles to always use
‘disable-warning-as-errors’ when ‘enable-native-coverage’ is used. It
should be pretty straightforward to do.
as you propose not to fix false positive warnings but fix make files instead.
Because warning
Hi,
my claim was based on the warnings which we were getting when we just
introduced code coverage builds in JDK 9, e.g. 8130790[1] (clobbered warning in
libt2k). these warnings haven't been seen w/o code coverage enabled, and
enabling coverage changes code path, so I don't think we should care
Hello,
On 2018-08-30 02:22, Magnus Ihse Bursie wrote:
On 2018-08-24 00:44, Igor Ignatev wrote:
Hi Leonid,
We have never supported native code coverage builds with warnings
enabled as errors. There are bugs in gcc which cause false positive
warnings, so it was decided to ignore all warnings
On 2018-08-24 00:44, Igor Ignatev wrote:
Hi Leonid,
We have never supported native code coverage builds with warnings enabled as
errors. There are bugs in gcc which cause false positive warnings, so it was
decided to ignore all warnings from instrumented builds. It’d be much better
and rel
Hi
Thanks for background. Assuming these or similar issues were already discussed
and it was decided to ignore warnings for code coverage builds I think it is
fine to fix build. Current warnings don't look as real issues for me also.
I moved components alias as bcc since now it is build-only c
Hi Leonid,
We have never supported native code coverage builds with warnings enabled as
errors. There are bugs in gcc which cause false positive warnings, so it was
decided to ignore all warnings from instrumented builds. It’d be much better
and reliable to fix makefiles to always use ‘disable-
