... and it had to be "fixed" 3-4 times by H.J.Lu before it went so
obscure to only hit people like me.
I run a weather forecasting system 4 times daily to test it out.
Because GCC would-be-4.4 is in regression fixes only during the last
four months, I thought I could use it with impunity.
Unfortunately, not so.
From the 17th of February onwards, I find the following diagnostics in
my output files:
-------------------------------
ABS(DPS)/3H = 1.007571 MB
MAXWIND = 70.835266 M/S IN (X,Y,LEV)= 129 145 15
STPS = 10077.023437500000
STQ = 4.921556472778
STS = 0.010799089447
STPE = 2497269760.000000000000
STKE = 1622026.375000000000
STTE = 2498891776.000000000000
TOTAL RAIN RATE(mm/d)= +Infinity
STRATIFORM PRECIP RATE= +Infinity
CONVECTIVE PRECIP RATE= 0.000000000000
CWPATH = +Infinity
COV2D = +Infinity
HIGH CLOUDS = +Infinity
MEDIUM CLOUDS = +Infinity
LOW CLOUDS = +Infinity
LOWEST LEVEL CLOUDS/FOG = +Infinity
AVERAGED CLOUD BASE= +Infinity
AVERAGED CLOUD TOP= +Infinity
SENF = -Infinity
LATF = NaN
MOMF = +Infinity
Without going into meteorological details, I hope you will trust me that
the Infinity's, NaN's should not be there.
After reverting r144098 and its updates today, the output looks normal:
-------------------------------
ABS(DPS)/3H = 1.144242 MB
MAXWIND = 75.913940 M/S IN (X,Y,LEV)= 155 151 13
STPS = 10076.395507812500
STQ = 4.889474391937
STS = 0.006965990644
STPE = 2498901504.000000000000
STKE = 1634238.125000000000
STTE = 2500535808.000000000000
TOTAL RAIN RATE(mm/d)= 0.150563895702
STRATIFORM PRECIP RATE= 0.146368727088
CONVECTIVE PRECIP RATE= 0.004195173737
CWPATH = 0.009085088037
COV2D = 0.115943998098
HIGH CLOUDS = 0.052225619555
MEDIUM CLOUDS = 0.031857494265
LOW CLOUDS = 0.062132436782
LOWEST LEVEL CLOUDS/FOG = 0.006665431429
AVERAGED CLOUD BASE= 448.916229248047
AVERAGED CLOUD TOP= 900.125244140625
SENF = -30.753757476807
LATF = -70.874824523926
MOMF = 0.177236527205
So *during phase 4 - regression fixes only* a change *introducing* two
peephole optimizations was not only proposed, but accepted and committed
(and "fixed", at least three times).
Unless a very good reason is in my inbox in the next 48 hours, I'll
revert this change under the obvious rule.
The GCC community has rules to guide its development process. These
rules are there so that third parties can evaluate our quality control.
We cannot and will not renege on them, however much we like the
particular developer who came up with this change.
Kind regards,
--
Toon Moene - e-mail: t...@moene.org (*NEW*) - phone: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
At home: http://moene.org/~toon/
Progress of GNU Fortran: http://gcc.gnu.org/gcc-4.4/changes.html
