This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
[offtopic] some statistics on testsuite-reports
- To: egcs at cygnus dot com
- Subject: [offtopic] some statistics on testsuite-reports
- From: Marc Lehmann <pcg at goof dot com>
- Date: Mon, 2 Nov 1998 02:25:23 +0100
Since I had no benchmark results to play with I added all testsuite reports
from my mail archive to my sql database. I got about 170 reports (I used to
remove testsuite results since they were big and used much space in the
folders, now I'm too lazy to do that. Also my script didn't parse all
subject lines, so this is not too representative).
Here are some fun results, I hope you find them a bit amusing (and only that).
"Who is sending most reports?" (who is really testing egcs?)
select count(testreport.id),submitter.name,submitter.address
from submitter,testreport where submitter.id=testreport.submitter
group by submitter.id order by 1 desc
+----------------------+-----------------------------+--------------------------------------+
| count(testreport.id) | name | address |
+----------------------+-----------------------------+--------------------------------------+
| 35 | Alexandre Oliva | oliva@dcc.unicamp.br |
| 17 | | manfred@s-direktnet.de |
| 15 | Mike Stump | mrs@wrs.com |
| 14 | David W. Schuler | schuld@btv.ibm.com |
| 11 | Robert Lipe | robertl@dgii.com |
| 9 | Billinghurst, David \(RTD\) | David.Billinghurst@riotinto.com.au |
| 8 | Horst von Brand | vonbrand@inf.utfsm.cl |
| 7 | David Rees | drees@oto.dyn.ml.org |
| 7 | Mumit Khan | khan@xraylith.wisc.edu |
| 7 | Nix | nix-egcs@esperi.demon.co.uk |
| 6 | Matthias Mueller | matthias@ica1.uni-stuttgart.de |
...
"Which systems report most failures?" (yeah!)
select count(sys.id),sys.sys from testresult,sys,testreport
where testresult.report=testreport.id and testreport.sys=sys.id
group by sys.id order by 1 desc
+---------------+---------------------------+
| count(sys.id) | sys |
+---------------+---------------------------+
| 1214 | i686-pc-sco3.2v5.0.5 |
| 1173 | mips-sgi-irix6.2 |
| 748 | i586-pc-linux-gnu |
| 704 | i586-pc-linux-gnulibc1 |
| 511 | sparc-sun-solaris2.5.1 |
| 399 | sparc-sun-sunos4.1.3 |
...
"Who reports most failures?" (robert seems to test less stable systems than mike!)
select count(submitter.id),submitter.name,submitter.address
from testreport,testresult,submitter where testresult.report=testreport.id
and submitter.id=testreport.submitter group by submitter.id order by 1 desc
+---------------------+-----------------------------+--------------------------------------+
| count(submitter.id) | name | address |
+---------------------+-----------------------------+--------------------------------------+
| 1456 | Alexandre Oliva | oliva@dcc.unicamp.br |
| 1214 | Robert Lipe | robertl@dgii.com |
| 1173 | Billinghurst, David \(RTD\) | David.Billinghurst@riotinto.com.au |
| 437 | | manfred@s-direktnet.de |
| 375 | David W. Schuler | schuld@btv.ibm.com |
| 362 | David Rees | drees@oto.dyn.ml.org |
| 354 | Mike Stump | mrs@wrs.com |
| 260 | Horst von Brand | vonbrand@inf.utfsm.cl |
| 243 | Matthias Mueller | matthias@ica1.uni-stuttgart.de |
| 230 | Theodore Papadopoulo | Theodore.Papadopoulo@sophia.inria.fr |
| 226 | Christian Joensson FOA 72 | chj@lin.foa.se |
...
"Which tests failed most often?"
select count(test.id),test.name,cause.message from testresult,test,cause
where testresult.test=test.id and testresult.cause=cause.id
group by test.id order by 1 desc limit 12
+----------------+-----------------------------------------+---------------------------------------------------------------+
| count(test.id) | name | message |
+----------------+-----------------------------------------+---------------------------------------------------------------+
| 544 | gcc.c-torture/execute/ieee/980619-1.c | FAIL: execution, -O2 -fomit-frame-pointer -finline-functions |
| 324 | gcc.c-torture/execute/980506-2.c | FAIL: execution, -O2 |
| 280 | gcc.c-torture/execute/loop-2g.c | FAIL: execution, -O2 -fomit-frame-pointer -finline-functions |
| 275 | gcc.c-torture/execute/loop-2f.c | FAIL: execution, -O2 -fomit-frame-pointer -finline-functions |
| 162 | gcc.c-torture/execute/980526-1.c | FAIL: execution, -O2 -fomit-frame-pointer -finline-functions |
| 152 | gcc.c-torture/execute/complex-5.c | FAIL: execution, -O2 -fomit-frame-pointer -finline-functions |
| 103 | g77.f-torture/execute/u77-test.f | FAIL: execution, -O2 -fomit-frame-pointer -finline-functions |
| 102 | gcc.c-torture/execute/931004-12.c | FAIL: execution, -O2 -fomit-frame-pointer -finline-functions |
| 99 | gcc.c-torture/execute/memcheck/blkarg.c | FAIL: execution, -O2 -fomit-frame-pointer -finline-functions |
| 90 | gcc.dg/980626-1.c | FAIL: (test for excess errors) |
| 90 | g77.f-torture/execute/971102-1.f | FAIL: compilation, -O0 |
| 80 | g++.robertl/eb132.C | XPASS: (test for excess errors) |
+----------------+-----------------------------------------+---------------------------------------------------------------+
Please don't tell me that my sql is broken - I know. And, if somebody
responsible is still reading: what SQL-Database could we use on the
egcs-webserver, and who writes the web-frontend? Open questions, open
questions....
-----==- |
----==-- _ |
---==---(_)__ __ ____ __ Marc Lehmann +--
--==---/ / _ \/ // /\ \/ / pcg@goof.com |e|
-=====/_/_//_/\_,_/ /_/\_\ --+
The choice of a GNU generation |
|