public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Jan Kratochvil <jan.kratochvil@redhat.com>
To: Joel Brobecker <brobecker@adacore.com>
Cc: gdb-patches@sourceware.org
Subject: Re: ping: [patch 5/6] testsuite: Fix prelink-support.exp without prelink installed
Date: Sun, 04 Jul 2010 10:19:00 -0000	[thread overview]
Message-ID: <20100704101842.GE6875@host0.dyn.jankratochvil.net> (raw)
In-Reply-To: <20100629220904.GV2595@adacore.com>

On Wed, 30 Jun 2010 00:09:04 +0200, Joel Brobecker wrote:
> I'm just concerned about matching the output
> from the prelink command, since it is assuming that it is going to
> be in English. Do you know if we force the language in the testsuite?

GDB testsuite does not force English, neither dejagnu.  But TCL seems to
ignore locale:
$ LANG=cs_CZ LC_ALL=cs_CZ tclsh
% exec /foo
couldn't execute "/foo": no such file or directory

Which may match:
3/30/99 (bug fix) Tcl no longer calls setlocale().  It looks at
env(LANG) and env(LC_TYPE) instead.  (stanton)

And some tcl grepping of sources I have done now.  (tcl8.5.8-src.tar.gz)
        sprintf(errSpace, "%dcouldn't execute \"%.150s\": ", errno, argv[0]);
 - no _() / gettext() there.


> I'm just wondering whether a more straightforward/brutal approach might
> be simpler and as effective: What if we searched for prelink in the PATH
> and skipped the entire test if not found?

There is explicitly executed /usr/sbin/prelink because on some systems
/usr/sbin is not in the default non-root $PATH.  This also fixes:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=225783
http://cvs.fedoraproject.org/viewvc/rpms/gdb/F-13/gdb-6.6-bz225783-prelink-path.patch?content-type=text%2Fplain&view=co
= gdb.base/prelink.exp: prelink -> /usr/sbin/prelink

It is true I did not ask if it is cross-OS compatible enough (Debian is OK).


Do you suggest /usr/sbin/prelink existence check, $PATH search (with
explicitly appended `/usr/sbin') or the current error message check?


Thanks,
Jan


gdb/testsuite/
2010-07-04  Jan Kratochvil  <jan.kratochvil@redhat.com>
	    Joel Brobecker  <brobecker@adacore.com>

	Cope with missing /usr/sbin/prelink.
	* lib/prelink-support.exp (prelink_no):
	<result == 1 && $output is "no such file or directory">: New.
	(prelink_yes): Likewise.  Return on failed prelink_no.

--- a/gdb/testsuite/lib/prelink-support.exp
+++ b/gdb/testsuite/lib/prelink-support.exp
@@ -187,6 +187,30 @@ proc prelink_no {arg {name {}}} {
     set result [catch $command output]
     verbose -log "result is $result"
     verbose -log "output is $output"
+    if {$result == 1 && [regexp {^(couldn't execute "/usr/sbin/prelink[^\r\n]*": no such file or directory\n?)*$} $output]} {
+	# Without prelink, at least verify that all the binaries do not
+	# contain the  ".gnu.prelink_undo" section (which would mean that they
+	# have already been prelinked).
+	set test "$test (missing /usr/sbin/prelink)"
+	foreach bin [split $arg] {
+	    if [string match "-*" $bin] {
+		# Skip prelink options.
+		continue
+	    }
+	    set readelf_program [transform readelf]
+	    set command "exec $readelf_program -WS $bin"
+	    verbose -log "command is $command"
+	    set result [catch $command output]
+	    verbose -log "result is $result"
+	    verbose -log "output is $output"
+	    if {$result != 0 || [string match {* .gnu.prelink_undo *} $output]} {
+		fail "$test ($bin is already prelinked)"
+		return 0
+	    }
+	}
+	pass $test
+	return 1
+    }
     if {$result == 0 && $output == ""} {
 	verbose -log "$name has been now unprelinked"
 	set command "exec /usr/sbin/prelink -uN $arg"
@@ -217,7 +241,9 @@ proc prelink_yes {arg {name ""}} {
     # Try to unprelink it first so that, if it has been already prelinked
     # before, we get a different address now, making the new result unaffected
     # by any previous prelinking.
-    prelink_no $arg "$name pre-unprelink"
+    if ![prelink_no $arg "$name pre-unprelink"] {
+	return 0
+    }
 
     set test "prelink $name"
 
@@ -239,6 +265,19 @@ proc prelink_yes {arg {name ""}} {
     set result [catch $command output]
     verbose -log "result is $result"
     verbose -log "output is $output"
+    if {$result == 1 && [regexp {^(couldn't execute "/usr/sbin/prelink[^\r\n]*": no such file or directory\n?)*$} $output]} {
+	set test "$test (missing /usr/sbin/prelink)"
+
+	# We could not find prelink.  We could check whether $args is already
+	# prelinked but we don't, because:
+	#   - It is unlikely that someone uninstalls prelink after having
+	#     prelinked the system ld.so;
+	#   - We still cannot change its prelinked address.
+	# Therefore, we just skip the test.
+
+	xfail $test
+	return 0
+    }
     if {$result == 0 && $output == ""} {
 	pass $test
 	return 1

  reply	other threads:[~2010-07-04 10:19 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-03-29 16:18 Jan Kratochvil
2010-06-09 15:10 ` ping: " Jan Kratochvil
2010-06-29 22:09   ` Joel Brobecker
2010-07-04 10:19     ` Jan Kratochvil [this message]
2010-07-05 17:40       ` Joel Brobecker
2010-07-05 18:13         ` Jan Kratochvil
2010-07-19 15:06       ` Pedro Alves
2010-07-19 15:43         ` Daniel Jacobowitz
2010-07-28 14:26           ` Disable i18n when running the testsuite (Re: ping: [patch 5/6] testsuite: Fix prelink-support.exp without prelink installed) Pedro Alves
2010-07-28 17:36             ` Joel Brobecker
2010-07-29 14:20               ` Disable i18n when running the testsuite Pedro Alves

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20100704101842.GE6875@host0.dyn.jankratochvil.net \
    --to=jan.kratochvil@redhat.com \
    --cc=brobecker@adacore.com \
    --cc=gdb-patches@sourceware.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).