Broken link to documentation

[Simon Wright]

In http://gcc.gnu.org/onlinedocs/, the Current development / User- 
Level Documentation link to GNAT User's Guide is broken.

Re: License compliance on updating gcc runtime libraries

[Simon Wright]

On 27 Feb 2019, at 12:41, Richard Kenner  wrote:
> 
> Remember that, from the perspective of copyright law, executing a program is 
> making a "copy"
> of that program.

Has that (rather extreme) view been litigated?

Re: License compliance on updating gcc runtime libraries

[Simon Wright]

On 27 Feb 2019, at 18:37, Richard Kenner  wrote:
> 
> 1) Whether executing a program is considered making a copy under
> copyright law.

I had a look through some of the published judgements, and it's clear that in 
the US at least copying into RAM (for whatever purpose, and provided the copy 
has more than fleeting duration) counts as copying for copyright purposes.

I can't help feeling (however useless feelings are in the face of the 9th 
Circuit) that a software publisher should be able to rely on other methods to 
retain control of their proprietary software. After all, ARM-based MCUs are 
capable of executing code directly from Flash, with no copying at all.

Release criteria for Darwin

[Simon Wright]

I see that, in the GCC 7 Release Criteria, the Secondary Platforms list 
includes i686-apple-darwin.

Should this now be x86_64-apple-darwin? I've been building this since GCC 
4.5.0, Darwin 10, in 2011.

'make check' questions

[Simon Wright]

I see from https://gcc.gnu.org/install/test.html that it's possible to run 
tests in parallel. I get the impression from gcc/Makefile that the check 
concerned has to be set up in the Makefile (in my build tree, configured with 
--target=x86_64-apple-darwin16 
--enable-languages=c,c++,ada,fortran,objc,obj-c++ , I see both lang_checks and 
lang_checks_parallelized set empty). So, is it necessary for check-ada or 
check-acats to cope with being run in parallel (i.e., will they ever see 
GCC_RUNTEST_PARALLELIZE_DIR set?)

Also in https://gcc.gnu.org/install/test.html, under what circumstances would a 
test report ERROR (the testsuite detected an error) or WARNING (the testsuite 
detected a possible problem)? For example, if a particular test that should 
compile & run has a build error, is that a FAIL or an ERROR? 

Re: 'make check' questions

[Simon Wright]

On 11 May 2017, at 11:50, Kyrill Tkachov  wrote:

> On 11/05/17 11:43, Simon Wright wrote:
>> I see from https://gcc.gnu.org/install/test.html that it's possible to run 
>> tests in parallel. I get the impression from gcc/Makefile that the check 
>> concerned has to be set up in the Makefile (in my build tree, configured 
>> with --target=x86_64-apple-darwin16 
>> --enable-languages=c,c++,ada,fortran,objc,obj-c++ , I see both lang_checks 
>> and lang_checks_parallelized set empty). So, is it necessary for check-ada 
>> or check-acats to cope with being run in parallel (i.e., will they ever see 
>> GCC_RUNTEST_PARALLELIZE_DIR set?)
> 
> I don't usually build Ada, but testing with "make -j check" works for me 
> where  is the parallelism I want

So it does. Will have to review the results to see how it's done in the Ada 
test scripts - deeply unclear to me at the moment.

>> Also in https://gcc.gnu.org/install/test.html, under what circumstances 
>> would a test report ERROR (the testsuite detected an error) or WARNING (the 
>> testsuite detected a possible problem)? For example, if a particular test 
>> that should compile & run has a build error, is that a FAIL or an ERROR?
> 
> ERROR results are usually problems with the testsuite infrastructure, like 
> misformed DejaGNU directives. They don't usually appear in a clean test run.
> If a test fails to build due to a compiler problem i.e. an ICE or other bug 
> it will be a FAIL. If the test harness has a problem with the testsuite 
> directives
> syntax I think it will be reported as ERROR.
> 
> I usually get WARNINGs if a runtime tests times out. It can happen when 
> testing against simulators or if the test was miscompiled into an infinite 
> loop.

OK, I see, thanks.

Re: timeouts/malloc failures in ada tests?

[Simon Wright]

On 7 Jul 2017, at 22:31, Martin Sebor  wrote:
> 
> I see large numbers of timeouts in Ada tests on trunk in parallel
> run s (make -j96) on x86_64.  Messages like the one below appear
> in the logs, suggesting some sort of heap corruption.  I'm having
> trouble reproducing it outside the rest of the test suite (i.e.,
> by just running the Ada tests by themselves) but maybe I'm doing
> it wrong.  Unless this is a known problem I can try to pinpoint
> it closer if someone could share the magic spell to run just Ada
> tests to speed up the debugging.

make check-gnat

check-ada runs the acats tests and then the gnat tests
check-acats runs just the acats tests
check-gnat runs just the gnat tests (and your problem is in the gnat tests; the 
acats tests don't use gnatclean)

Re: GCC interpretation of C11 atomics (DR 459)

[Simon Wright]

On 27 Feb 2018, at 12:56, Ruslan Nikolaev via gcc  wrote:
> 
> And all this mess to accommodate almost non-existent case when someone wants 
> to use atomic_load on read-only memory for wide types, in which no good 
> solution exists anyway

Sorry to butt in, but - if it's ROM why would you need atomic load anyway? (of 
course, if it's just a constant view of the object, reason is obvious)

Re: GCC interpretation of C11 atomics (DR 459)

[Simon Wright]

On 27 Feb 2018, at 17:07, Torvald Riegel  wrote:
> 
> On Tue, 2018-02-27 at 16:40 +, Simon Wright wrote:
>> On 27 Feb 2018, at 12:56, Ruslan Nikolaev via gcc  wrote:
>>> 
>>> And all this mess to accommodate almost non-existent case when someone 
>>> wants to use atomic_load on read-only memory for wide types, in which no 
>>> good solution exists anyway
>> 
>> Sorry to butt in, but - if it's ROM why would you need atomic load anyway? 
>> (of course, if it's just a constant view of the object, reason is obvious)
> 
> Consider a producer-consumer relationship between two processes where
> the producer doesn't want to wait for the consumer.  For example, the
> producer could be an application that's being traced, and the consumer
> is a trace aggregation tool.  The producer can provide a read-only
> mapping to the consumer, and put a nonblocking ring buffer or something
> similar in there.  That allows the consumer to read, but it still needs
> atomic access because the consumer is modifying the ring buffer
> concurrently.

OK, got that, thanks (this is what I meant by "just a constant view of the 
object", btw).

Misled by "read-only memory" since in the embedded world ROM (usually actually 
in Flash) is effectively read-only to all.

Re: interrupting Ada tests

[Simon Wright]

On 5 Jul 2018, at 20:08, Richard Biener  wrote:
> 
> On July 5, 2018 6:37:58 PM GMT+02:00, Martin Sebor  wrote:
>> Ada tests don't seem to respond to the INT signal: when
>> I interrupt a parallel make check while the Ada tests are
>> running, other test suites are interrupted as well and go
>> away, but ada tests keep running.  Is there some trick to
>> have Ctrl-C have the expected effect on the Ada test suite
>> as well or is this is a bug?
> 
> It only happens sometimes to me. I guess this is a bug. 
> 
> Richard. 

I'm pretty sure it's a bug, in the 'acats' part of the test, not the 'gnat' 
part; it's happened to me repeatedly while developing an upgraded ACATS suite 
over at https://github.com/simonjwright/ACATS 
.

(by the way, to run just this part of the suite, it's 'make check-acats').

No doubt the problem is that the suite runs as a shell script, and the C-c 
doesn't get applied at the right place(s).

I'm not sure that I'm up to what would be a major recasting of the 2,500 or so 
test programs into a dejagnu framework (would that be likely to fix this 
issue?).

>> Thanks
>> Martin
> 

Mercurial mirror

[Simon Wright]

The Mercurial mirror at http://gcc.gnu.org/hg/gcc was last updated 11 months 
ago at SVN r166522.

I think it can only cause confusion to have the mirror live but stale; ought it 
to be turned off?

Mercurial mirror

[Simon Wright]

The Mercurial mirror at http://gcc.gnu.org/hg/gcc was last updated 11 months 
ago at SVN r166522.

I think it can only cause confusion to have the mirror live but stale; ought it 
to be turned off?