Test suite regressions due to integer-gmp2

According to my performance builders, changeset:c774b28f/ghc (Phab:D82) caused prof-doc-fib and linker_unload to fail on both performance builders (Ubuntu 13.10 and Ubuntu 14.04):

Wrong exit code (expected 0 , actual 2 )
Stdout:

Stderr:
linker_unload: /home/nomeata/logs/ghc-tmp-REV/libraries/integer-gmp2/dist-install/build/libHSinteg_21cuTlnn00eFNd4GMrxOMi.a: unknown symbol `__gmpn_rshift'
linker_unload: resolveObjs failed
make[3]: *** [linker_unload] Error 1

*** unexpected failure for linker_unload(normal)

and

Actual prof output differs from expected:
--- ./profiling/should_run/prof-doc-fib.prof.sample     2014-12-01 15:30:19.000000000 +0100
+++ ./profiling/should_run/prof-doc-fib.prof    2014-12-01 15:56:08.000000000 +0100
@@ -1,9 +1,9 @@
-       Thu Oct 27 09:29 2011 Time and Allocation Profiling Report  (Final)
+       Mon Dec  1 15:56 2014 Time and Allocation Profiling Report  (Final)
 
-          fib +RTS -p -RTS
+          prof-doc-fib +RTS -hc -p -RTS
 
-       total time  =        0.76 secs   (38 ticks @ 20 ms)
-       total alloc = 247,940,020 bytes  (excludes profiling overheads)
+       total time  =        0.14 secs   (135 ticks @ 1000 us, 1 processor)
+       total alloc = 107,829,304 bytes  (excludes profiling overheads)
 
 COST CENTRE MODULE  %time %alloc
 
@@ -13,13 +13,16 @@
                                                       individual     inherited
 COST CENTRE MODULE                  no.     entries  %time %alloc   %time %alloc
 
-MAIN        MAIN                    102           0    0.0    0.0   100.0  100.0
- CAF        Main                    203           0    0.0    0.0   100.0  100.0
-  main      Main                    204           1    0.0    0.0   100.0  100.0
-   main.g   Main                    207           1    0.0    0.0     0.0    0.1
-    fib     Main                    208        1973    0.0    0.1     0.0    0.1
-   main.f   Main                    205           1    0.0    0.0   100.0   99.9
-    fib     Main                    206     2692537  100.0   99.9   100.0   99.9
- CAF        GHC.Conc.Signal         201           0    0.0    0.0     0.0    0.0
- CAF        GHC.IO.Encoding.Iconv   191           0    0.0    0.0     0.0    0.0
- CAF        GHC.IO.Handle.FD        183           0    0.0    0.0     0.0    0.0
+MAIN        MAIN                     45           0    0.0    0.0   100.0  100.0
+ main       Main                     91           0    0.0    0.0     0.0    0.0
+ CAF        Main                     89           0    0.0    0.0   100.0  100.0
+  main      Main                     90           1    0.0    0.0   100.0  100.0
+   main.f   Main                     94           1    0.0    0.0   100.0   99.9
+    fib     Main                     95     2692537  100.0   99.9   100.0   99.9
+   main.g   Main                     92           1    0.0    0.0     0.0    0.1
+    fib     Main                     93        1973    0.0    0.1     0.0    0.1
+ CAF        GHC.IO.Handle.Text       86           0    0.0    0.0     0.0    0.0
+ CAF        GHC.IO.Handle.FD         82           0    0.0    0.0     0.0    0.0
+ CAF        GHC.Conc.Signal          78           0    0.0    0.0     0.0    0.0
+ CAF        GHC.IO.Encoding          76           0    0.0    0.0     0.0    0.0
+ CAF        GHC.IO.Encoding.Iconv    75           0    0.0    0.0     0.0    0.0
*** unexpected failure for prof-doc-fib(profasm)

The former is also observed by SPJ. The latter actually looks less like a regression, and more an improvement – maybe Herbert simply did not run a profiled version when updating test results?

Trac metadata

Trac field	Value
Version	7.9
Type	Bug
TypeOfFailure	OtherFailure
Priority	normal
Resolution	Unresolved
Component	Compiler
Test case
Differential revisions
BlockedBy
Related
Blocking
CC	hvr
Operating system
Architecture

changed weight to 5

added Tbug Trac import labels

Replying to nomeata:

maybe Herbert simply did not run a profiled version when updating test results?

1. ..if ./validate doesn't run a profiled version, then I probably never ran the tests w/ a profiled version... :-/

How do I perform a profiled testsuite run?

I’m never sure about the zoo of possible combinations. ghcspeed, for instance, has these settings:

echo 'GhcLibWays := $(filter v dyn,$(GhcLibWays))' >> mk/build.mk
echo 'GhcLibHcOpts += -O -dcore-lint'  >> mk/build.mk
echo 'GhcStage2HcOpts += -O -dcore-lint'  >> mk/build.mk

which tires to match the validate settings, but I notice it leaves out p (which validate does not). It also only calls make -C testsuite fast.

This leaves me confused how it could run prof-doc-fib at all. Maybe it’s because I run perl boot before creating mk/build.mk?

Simon Marlow say: the resolveObjs problem is not a bug in GHC; it's a bug in the test. But it's not easy to fix.

changed milestone to %7.10.1

assigned to @simonmar

Trac metadata

Trac field	Value
Component	Compiler → Test Suite

added testsuite label

closed

Should now work: 6108d95a

Trac metadata

Trac field	Value
Resolution	Unresolved → ResolvedFixed

is this to be merged into 7.10?

Yes please!

reopened

unassigned @simonmar

Unfortunately, it did not fix it on my build machine:

Stdout:

Stderr:
linker_unload: /data1/ghc-builder/logs/ghc-tmp-REV/libraries/integer-gmp2/dist-install/build/libHSinteg_6zeGtnFHpaVBJ80QaL9uVu.a: unknown symbol `__gmpn_rshift'
linker_unload: resolveObjs failed
make[3]: *** [linker_unload] Fehler 1

*** unexpected failure for linker_unload(normal)

Trac metadata

Trac field	Value
Resolution	ResolvedFixed → Unresolved

Sorry, got confused and were looking at an old log.

added 1 deleted label

closed

removed 1 deleted label

Merged to ghc-7.10 (via 7f9704f4).

Trac metadata

Trac field	Value
Resolution	Unresolved → ResolvedFixed

added Pnormal label