This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: test results web page


 =>From: Jeffrey A Law <law@hurl.cygnus.com>
 =>...
 =>How do I submit results to the database?

Mail them to egcs-results@sweet.webvision.com.  The parser is pretty
stupid, though, so if you don't send the output of Alexandre's script,
it probably won't go into the DB.  Also, it doesn't bother to tell you
what was wrong if it fails.

 =>  > The data are stored in a PostgreSQL database running on a Sun; right
 =>Hmmm, so what software would we need if we wanted to run this off
 =>the egcs web server?  That is unless you want to host this database
 =>and we'll just link to your server from the egcs home page.

I can run it here, but if you think people are going to hack on it,
perhaps a more accessible location would be better.  You need
PostgreSQL, perl+DBI, and a standard web server.  I'm running
Netscape, but you could use Apache.

 =>Not everyone is likely to run the testsuite; it would be nice to
 =>be able to record that someone successfully built a particular
 =>version on a particular target.  Maybe use the file upload facility
 =>to send config.guess and version.c to get this info?  Or do we have
 =>them type it in?  And (of course) a comments area for these folks.

Good idea...I'll add that.

 =>From: Joe Buck <jbuck@Synopsys.COM>
 =>...
 =>If I understand right, he's storing the output of Alexandre's script
 =>(e.g. only the FAILs and XPASSes), not the whole summary.

FAILs and XPASSes are *not* saved now; only the "# of expected
passes," etc.  I'll see what I can do for the individual test lines.
You're right, though, that I only deal with the output of Alexandre's
script, and not the raw test result files.

 =>While I know it's tricky, I'd like to be able to do queries on individual
 =>failures: e.g. show all tests that FAIL on sparc-* that do not FAIL on
 =>ix86-* and the like.

That's theoretically possible, but I don't know whether PostgreSQL is
up to it.  More stuff to check out...

d.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]