Re: [RFA]: Fix for pending breakpoints in manually loaded/unloaded shlibs

[Michael Chastain <mec dot gnu at mindspring dot com>]

I ran it on native hppa2.0w-hp-hpux11.11 and it barfed.

  Running /house/chastain/gdb/s1/gdb/testsuite/gdb.base/unload.exp ...
  gdb compile failed, /usr/ccs/bin/ld: Can't find library: "dl"
  WARNING: Testcase compile failed, so all tests in this file will automatically fail.

  ERROR: (timeout) GDB never initialized after 10 seconds.
  WARNING: remote_expect statement without a default case?!
  ... and so on ...

This is with hp ansi c B.11.11.28706.GP.

I looked at the hpux man page for dlopen and it looks similar to
the linux man page.  (And the linux man page says that their
interface comes from solaris).

The hpux man page for dlopen explicitly says "-ldl" but there is
no such library.

If I rip out the "-ldl" option" then the test script runs
and I get this:

  Running target unix
  Running /house/chastain/gdb/s1/gdb/testsuite/gdb.base/unload.exp ...
  PASS: gdb.base/unload.exp: set pending breakpoint
  PASS: gdb.base/unload.exp: single pending breakpoint info
  FAIL: gdb.base/unload.exp: running program
  FAIL: gdb.base/unload.exp: continuing to end of program
  FAIL: gdb.base/unload.exp: rerun to shared library breakpoint
  FAIL: gdb.base/unload.exp: continuing to end of program second time

Which is okay.  The test script is doing its job: making gdb ill.

I'll write up a little patch and send it to you.

Also I am working on a new "gdb_build" proc that will encapsulate all
this compiler-specific and target-specific mumbo-jumbo.  But for now
let's continue with keeping the logic in the test script, so you can
get the test script checked in.

Michael C