Merge back in some code changes made in the Ryu patch that really belong
here, in preparation for rebasing Ryu on top of this (since this is
really a separate functional change). Posting this mainly to let cfbot
take a look at it.
--
Andrew (irc:RhodiumToad)
diff --git a/configure b/configure
in
This one builds ok on appveyor with at least three different VS
versions. Though I've not tried the exact combination of commands
used by cfbot... yet.
--
Andrew (irc:RhodiumToad)
diff --git a/configure b/configure
index 06fc3c6835..e3176e24e9 100755
--- a/configure
+++ b/configure
@@ -15802,6 +
> "Andrew" == Andrew Gierth writes:
Andrew> Because it turns out that Windows (at least the version running
Andrew> on Appveyor) completely fucks this up; strtof() is apparently
Andrew> returning infinity or zero _without setting errno_ for values
Andrew> out of range for float: input of
> "Tom" == Tom Lane writes:
>> (FWIW, I'm running the patch on gaur's host, just to confirm it
>> does what you expect. Should have an answer in an hour or so ...)
Tom> It does --- it compiles cleanly, and the float4 output matches
Tom> float4-misrounded-input.out.
Well I'm glad _somet
I wrote:
> (FWIW, I'm running the patch on gaur's host, just to confirm it
> does what you expect. Should have an answer in an hour or so ...)
It does --- it compiles cleanly, and the float4 output matches
float4-misrounded-input.out.
(This is the v1 patch not v2.)
regar
> "Tom" == Tom Lane writes:
> Andrew Gierth writes:
>> See if Windows likes this one any better.
Tom> Doubtful, because AFAICT it's the same patch.
Sigh. copied wrong file. Let's try that again.
--
Andrew (irc:RhodiumToad)
diff --git a/configure b/configure
index 06fc3c6835..e3176e24
Andrew Gierth writes:
> See if Windows likes this one any better.
Doubtful, because AFAICT it's the same patch.
regards, tom lane
See if Windows likes this one any better.
--
Andrew (irc:RhodiumToad)
diff --git a/configure b/configure
index 06fc3c6835..e3176e24e9 100755
--- a/configure
+++ b/configure
@@ -15802,6 +15802,19 @@ esac
fi
+ac_fn_c_check_func "$LINENO" "strtof" "ac_cv_func_strtof"
+if test "x$ac_cv_func_str
Andrew Gierth writes:
> "Tom" == Tom Lane writes:
> Tom> The errno handling in strtof seems bizarrely overcomplex; why do
> Tom> you need the separate caller_errno variable?
> Eh. I was preserving the conventional behaviour of setting errno only on
> errors and not resetting it to 0,
Oh, I se
> "Tom" == Tom Lane writes:
>> As discussed in the Ryu thread, herewith a draft of a patch to use
>> strtof() for float4 input (rather than using strtod() with its
>> double-rounding issue).
Tom> The errno handling in strtof seems bizarrely overcomplex; why do
Tom> you need the separate
Andrew Gierth writes:
> As discussed in the Ryu thread, herewith a draft of a patch to use
> strtof() for float4 input (rather than using strtod() with its
> double-rounding issue).
The errno handling in strtof seems bizarrely overcomplex; why do
you need the separate caller_errno variable?
> Th
As discussed in the Ryu thread, herewith a draft of a patch to use
strtof() for float4 input (rather than using strtod() with its
double-rounding issue).
An exhaustive search shows that this does not change the resulting
bit-pattern for any input string that could have been generated by PG
with ex
12 matches
Mail list logo
