I mentioned in passing that Fedora 17 (currently in beta) ships
Lilypond 2.15.29 that reports bogus bar checks.
It turns out that both the current Lilypond from git and Lilypond
2.14.2 are miscompiled on the up-to-date Fedora 17 so they exhibit the
same problem.
What compiler version are
On Sat, 9 Jun 2012 21:19:31 +0200
Frédéric Bron frederic.b...@m4x.org wrote:
I just installed fedora 17 x86_64.
After update, it comes with g++ 4.7.0-5 and lilypon 2.15.39.
Please find attached the output which looks good to me.
Yes, it looks good.
Any idea why it does not come with a
On Sat, Jun 09, 2012 at 09:19:31PM +0200, Frédéric Bron wrote:
After update, it comes with g++ 4.7.0-5 and lilypon 2.15.39.
Please find attached the output which looks good to me.
Any idea why it does not come with a stable release (i.e. 2.14)?
Because the fedora mainter(s) for lilypond think
On Sat, Jun 9, 2012 at 4:58 PM, Graham Percival
gra...@percival-music.ca wrote:
I've been fairly clear: every release says either
It is strongly recommended that normal users do not use this
release, and instead use the stable 2.14 version.
or
All users are invited to experiment with this
On Sat, Jun 09, 2012 at 07:55:30PM -0300, Han-Wen Nienhuys wrote:
On Sat, Jun 9, 2012 at 4:58 PM, Graham Percival
gra...@percival-music.ca wrote:
I've been fairly clear: every release says either
It is strongly recommended that normal users do not use this
release, and instead use the
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
We can probably use yes:400600[0-2]) here: 4.6.3 is supposed to contain
the fix.
Thank you for opening a separate ticket for it (2514) and reviewing my
ticket (2513)!
Well, it means that my ticket goes through
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
Now we just need to figure out which of the gazillion optimization
options (listed in the assembly file) from O2 as compared to O1 makes
the compilation go bad, and then we can add appropriate compiler
options.
We had
Quoting David Kastrup d...@gnu.org:
We can probably use yes:400600[0-2]) here: 4.6.3 is supposed to contain
the fix.
Thank you for opening a separate ticket for it (2514) and reviewing my
ticket (2513)!
--
Regards,
Pavel Roskin
___
Pavel Roskin pro...@gnu.org writes:
Quoting Pavel Roskin pro...@gnu.org:
Fedora 17 i386 has exactly the same problem.
Let me try to find the file that needs to be compiled with -O1. I have
some ideas how to bisect the problem. I want to write a wrapper around
g++ that starts adding -O1
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
Pavel Roskin pro...@gnu.org writes:
Quoting Pavel Roskin pro...@gnu.org:
I'll try to identify the function tomorrow.
That would be welcome. I looked through the function right now and saw
no obvious problematic
David Kastrup d...@gnu.org writes:
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
Pavel Roskin pro...@gnu.org writes:
Quoting Pavel Roskin pro...@gnu.org:
I'll try to identify the function tomorrow.
That would be welcome. I looked through the function right
David Kastrup d...@gnu.org writes:
The optimizer does not appear interested in the result of the
comparison when calculating the minimum value! There is no
conditional path entered. That does not make all too much sense.
Since you managed to boil this down to a single function and
Quoting David Kastrup d...@gnu.org:
That would have been my most likely guess. Probably the min function
call.
You are absolultely right! I was able to work around the gcc bug by
making a separate function for Moment comparison and using it instead
of min():
Moment
min_moment (Moment
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
That would have been my most likely guess. Probably the min function
call.
You are absolultely right! I was able to work around the gcc bug by
making a separate function for Moment comparison and using it instead
of
Quoting David Kastrup d...@gnu.org:
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
That would have been my most likely guess. Probably the min function
call.
You are absolultely right! I was able to work around the gcc bug by
making a separate function for
Quoting David Kastrup d...@gnu.org:
Now we just need to figure out which of the gazillion optimization
options (listed in the assembly file) from O2 as compared to O1 makes
the compilation go bad, and then we can add appropriate compiler
options.
We had to do this sort of $!# already for gcc
Quoting David Kastrup d...@gnu.org:
The optimizer does not appear interested in the result of the comparison
when calculating the minimum value! There is no conditional path
entered. That does not make all too much sense.
That's a truly sharp eye!
A diff between the i386 assembly without
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
That would have been my most likely guess. Probably the min function
call.
You are absolultely right! I was able to work around the gcc bug
Quoting David Kastrup d...@gnu.org:
Pavel Roskin pro...@gnu.org writes:
Hello!
I mentioned in passing that Fedora 17 (currently in beta) ships
Lilypond 2.15.29 that reports bogus bar checks.
It turns out that both the current Lilypond from git and Lilypond
2.14.2 are miscompiled on the
Pavel Roskin pro...@gnu.org writes:
Quoting David Kastrup d...@gnu.org:
Pavel Roskin pro...@gnu.org writes:
Hello!
I mentioned in passing that Fedora 17 (currently in beta) ships
Lilypond 2.15.29 that reports bogus bar checks.
It turns out that both the current Lilypond from git and
Quoting David Kastrup d...@gnu.org:
Any chance to compile using i386? If the problem persisted, it would
strongly point to the language frontend of g++ being involved. If it
vanishes, it can be either the backend or a general 32/64 bit
architecture problem.
Fedora 17 i386 has exactly the
Quoting Pavel Roskin pro...@gnu.org:
Fedora 17 i386 has exactly the same problem.
Let me try to find the file that needs to be compiled with -O1. I have
some ideas how to bisect the problem. I want to write a wrapper around
g++ that starts adding -O1 to the command line after N invocations.
Hello!
I mentioned in passing that Fedora 17 (currently in beta) ships
Lilypond 2.15.29 that reports bogus bar checks.
It turns out that both the current Lilypond from git and Lilypond
2.14.2 are miscompiled on the up-to-date Fedora 17 so they exhibit the
same problem.
That's the
Pavel Roskin pro...@gnu.org writes:
Hello!
I mentioned in passing that Fedora 17 (currently in beta) ships
Lilypond 2.15.29 that reports bogus bar checks.
It turns out that both the current Lilypond from git and Lilypond
2.14.2 are miscompiled on the up-to-date Fedora 17 so they exhibit
24 matches
Mail list logo