Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Jan Kratochvil <jan.kratochvil@redhat.com>
To: Doug Evans <dje@google.com>
Cc: gdb-patches@sourceware.org
Subject: [patchv2] [testsuite] Fix running in-src-dir in the same directory twice
Date: Fri, 29 Aug 2014 21:10:00 -0000	[thread overview]
Message-ID: <20140829211045.GA5890@host2.jankratochvil.net> (raw)
In-Reply-To: <20140829131326.GA19760@host2.jankratochvil.net>

[-- Attachment #1: Type: text/plain, Size: 1150 bytes --]

On Fri, 29 Aug 2014 15:13:26 +0200, Jan Kratochvil wrote:
> On Thu, 28 Aug 2014 23:53:59 +0200, Doug Evans wrote:
> > Tests can prematurely terminate for various unanticipatable reasons.
> > Can this be solved by making sure any possibly hazardous left
> > over files are cleaned up at start?
> 
> I think it would be best to modify the testcase so that the '. -> symlink'
> is made in a special subdirectory so that the *.exp files loop cannot occur
> even if the symlink is left there.

Done here.  Therefore it no longer tries to delete the symlink as it no longer
does any harm:

$ ls -l testsuite/gdb.base/argv0-symlink.*
-rw-r--r-- 1 jkratoch jkratoch  803 Aug 29 21:59 testsuite/gdb.base/argv0-symlink.c
-rw-r--r-- 1 jkratoch jkratoch 2839 Aug 29 21:59 testsuite/gdb.base/argv0-symlink.exp

testsuite/gdb.base/argv0-symlink.d:
total 16
-rwxr-xr-x 1 jkratoch jkratoch 9493 Aug 29 22:45 argv0-symlink*
lrwxrwxrwx 1 jkratoch jkratoch    1 Aug 29 22:45 argv0-symlink-dirlink -> ./
lrwxrwxrwx 1 jkratoch jkratoch   13 Aug 29 22:45 argv0-symlink-filelink -> argv0-symlink*
-rw-r--r-- 1 jkratoch jkratoch 2824 Aug 29 22:45 argv0-symlink0.o
$ _


Jan

[-- Attachment #2: subdir.patch --]
[-- Type: text/plain, Size: 2637 bytes --]

gdb/testsuite/
2014-08-29  Jan Kratochvil  <jan.kratochvil@redhat.com>

	Fix running in-src-dir in the same directory twice.
	* gdb.base/argv0-symlink.exp: Move everything into a subdirectory.

diff --git a/gdb/testsuite/gdb.base/argv0-symlink.exp b/gdb/testsuite/gdb.base/argv0-symlink.exp
index d849b4c..2005e7c 100644
--- a/gdb/testsuite/gdb.base/argv0-symlink.exp
+++ b/gdb/testsuite/gdb.base/argv0-symlink.exp
@@ -15,21 +15,26 @@
 
 standard_testfile
 
-if { [build_executable ${testfile}.exp ${testfile} ${srcfile}] == -1 } {
+set outputbase ${testfile}.d
+set outputdir [standard_output_file $outputbase]
+remote_exec host "rm -rf $outputdir"
+file mkdir $outputdir
+
+if { [build_executable ${testfile}.exp [file join $outputbase ${testfile}] ${srcfile}] == -1 } {
     return -1
 }
 
 set test "kept file symbolic link name"
 set filelink "${testfile}-filelink"
 
-remote_file host delete [standard_output_file $filelink]
-set status [remote_exec host "ln -sf ${testfile} [standard_output_file $filelink]"]
+remote_file host delete [file join $outputdir $filelink]
+set status [remote_exec host "ln -sf ${testfile} [file join $outputdir $filelink]"]
 if {[lindex $status 0] != 0} {
     unsupported "$test (host does not support symbolic links)"
     return 0
 }
 
-clean_restart "$filelink"
+clean_restart [file join $outputbase $filelink]
 
 if ![runto_main] {
     untested "could not run to main"
@@ -44,7 +49,7 @@ gdb_test {print argv[0]} "/$filelink\"" $test
 # For a link named /PATH/TO/DIR/LINK, we want to check the output
 # against "/DIR/LINK", but computed in a way that doesn't make
 # assumptions about the test directory layout.
-set full_filelink [standard_output_file $filelink]
+set full_filelink [file join $outputdir $filelink]
 set lastdir [file tail [file dirname $full_filelink]]
 
 gdb_test "info inferiors" "/$lastdir/$filelink *" "$test for info inferiors"
@@ -56,14 +61,14 @@ set dirlink "${testfile}-dirlink"
 # 'ln -sf' does not overwrite symbol link to a directory.
 # 'remote_file host delete' uses stat (not lstat), therefore it refuses to
 # delete a directory.
-remote_exec host "rm -f [standard_output_file $dirlink]"
-set status [remote_exec host "ln -sf . [standard_output_file $dirlink]"]
+remote_exec host "rm -f [file join $outputdir $dirlink]"
+set status [remote_exec host "ln -sf . [file join $outputdir $dirlink]"]
 if {[lindex $status 0] != 0} {
     unsupported "$test (host does not support symbolic links)"
     return 0
 }
 
-clean_restart "$dirlink/$filelink"
+clean_restart [file join $outputbase $dirlink $filelink]
 
 if ![runto_main] {
     untested "could not run to main"

      reply	other threads:[~2014-08-29 21:10 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-08-28 21:10 [patch] " Jan Kratochvil
2014-08-29 12:38 ` Doug Evans
2014-08-29 12:49   ` Jan Kratochvil
2014-08-29 13:13   ` Jan Kratochvil
2014-08-29 21:10     ` Jan Kratochvil [this message]

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=20140829211045.GA5890@host2.jankratochvil.net \
    --to=jan.kratochvil@redhat.com \
    --cc=dje@google.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