* libgcj-$VERSION changes
@ 2002-05-17 0:19 David S. Miller
2002-05-17 12:33 ` Anthony Green
0 siblings, 1 reply; 3+ messages in thread
From: David S. Miller @ 2002-05-17 0:19 UTC (permalink / raw)
To: java; +Cc: gcc-bugs, green
Anthony, why did you hard code the value for the version
number in the 3.1 branch version of your libjava.exp change:
2002-04-11 Anthony Green <green@redhat.com>
* lib/libjava.exp: Use libgcj-3.1.jar, not libgcj.jar.
Whereas on the mainline you used gcc_version which is really
how the libgcj-$VER number is determined by the build process?
2002-04-12 Anthony Green <green@redhat.com>
* lib/libjava.exp: Use libgcj-VERSION.jar, not libgcj.jar.
This is causing problems on the 3.1 branch now because gcc_version
is 3.1.1 but the testsuite is looking for libgcj-3.1.jar
So the whole libjava testsuite fails now because the classpath
is being set incorrectly.
It seems to me that your mainline version of the change should be put
into the 3.1 branch libjava.exp like so:
--- libjava/testsuite/lib/libjava.exp.~1~ Wed May 8 19:04:05 2002
+++ libjava/testsuite/lib/libjava.exp Fri May 17 00:05:28 2002
@@ -112,6 +112,7 @@
global TOOL_EXECUTABLE
global original_ld_library_path
global env objdir
+ global env gcc_version
global tool_root_dir
global libjava_libgcc_s_path
@@ -129,6 +130,11 @@
}
}
+ # Determine the version so we can find the libgcj jar file.
+ set text [eval exec "$GCJ_UNDER_TEST --version 2>@ stdout"]
+ regexp -- "gcj \[^ \]+ (\[^ \]+) .*" $text ignore gcc_version
+ verbose "jar file is libgcj-$gcc_version.jar"
+
# FIXME: This finds libgcj.spec for the default multilib.
# If thread models differ between multilibs, this has to be moved
# to libjava_arguments
@@ -265,6 +271,7 @@
global runtests
global env
global tool_root_dir
+ global gcc_version
global libjava_libgcc_s_path
if [info exists LIBJAVA] {
@@ -330,9 +337,9 @@
verbose "LD_LIBRARY_PATH = $env(LD_LIBRARY_PATH)"
# Set the CLASSPATH environment variable
- verbose "CLASSPATH is .:$srcdir/$subdir:$objdir:$objdir/../libgcj-3.1.jar"
+ verbose "CLASSPATH is .:$srcdir/$subdir:$objdir:$objdir/../libgcj-$gcc_version.jar"
global env
- set env(CLASSPATH) ".:$srcdir/$subdir:$objdir:$objdir/../libgcj-3.1.jar"
+ set env(CLASSPATH) ".:$srcdir/$subdir:$objdir:$objdir/../libgcj-$gcc_version.jar"
if {$mode == "link"} {
global wrapper_file wrap_compile_flags;
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: libgcj-$VERSION changes
2002-05-17 0:19 libgcj-$VERSION changes David S. Miller
@ 2002-05-17 12:33 ` Anthony Green
2002-05-17 14:12 ` David S. Miller
0 siblings, 1 reply; 3+ messages in thread
From: Anthony Green @ 2002-05-17 12:33 UTC (permalink / raw)
To: David S. Miller; +Cc: java, gcc-bugs
On Fri, 2002-05-17 at 00:05, David S. Miller wrote:
>
> Anthony, why did you hard code the value for the version
> number in the 3.1 branch version of your libjava.exp change:
Ah yes. Because at the time it seemed like 3.1 was about to go out
within a couple of days. Versioning the jar file was important for the
release, but I had no confidence that I would be able to create a proper
patch in time for the release -- so we decided that the hack would do.
As it happened, I did come up with a proper patch the next day, but only
applied it to the trunk for fear of destabilizing the branch.
> It seems to me that your mainline version of the change should be put
> into the 3.1 branch libjava.exp like so:
Yes -- I agree. I don't think any of us thought about the impact of the
branch patch post 3.1... :-(
Thanks for looking into this David.
AG
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: libgcj-$VERSION changes
2002-05-17 12:33 ` Anthony Green
@ 2002-05-17 14:12 ` David S. Miller
0 siblings, 0 replies; 3+ messages in thread
From: David S. Miller @ 2002-05-17 14:12 UTC (permalink / raw)
To: green; +Cc: java, java-patches
From: Anthony Green <green@redhat.com>
Date: 17 May 2002 12:32:53 -0700
On Fri, 2002-05-17 at 00:05, David S. Miller wrote:
> It seems to me that your mainline version of the change should be put
> into the 3.1 branch libjava.exp like so:
Yes -- I agree. I don't think any of us thought about the impact of the
branch patch post 3.1... :-(
Thanks for looking into this David.
No problem.
Installed on the branch as follows:
2002-04-12 Anthony Green <green@redhat.com>
* lib/libjava.exp: Use libgcj-VERSION.jar, not libgcj-3.1.jar.
--- testsuite/lib/libjava.exp.~1~ Wed May 8 19:04:05 2002
+++ testsuite/lib/libjava.exp Fri May 17 00:05:28 2002
@@ -112,6 +112,7 @@
global TOOL_EXECUTABLE
global original_ld_library_path
global env objdir
+ global env gcc_version
global tool_root_dir
global libjava_libgcc_s_path
@@ -129,6 +130,11 @@
}
}
+ # Determine the version so we can find the libgcj jar file.
+ set text [eval exec "$GCJ_UNDER_TEST --version 2>@ stdout"]
+ regexp -- "gcj \[^ \]+ (\[^ \]+) .*" $text ignore gcc_version
+ verbose "jar file is libgcj-$gcc_version.jar"
+
# FIXME: This finds libgcj.spec for the default multilib.
# If thread models differ between multilibs, this has to be moved
# to libjava_arguments
@@ -265,6 +271,7 @@
global runtests
global env
global tool_root_dir
+ global gcc_version
global libjava_libgcc_s_path
if [info exists LIBJAVA] {
@@ -330,9 +337,9 @@
verbose "LD_LIBRARY_PATH = $env(LD_LIBRARY_PATH)"
# Set the CLASSPATH environment variable
- verbose "CLASSPATH is .:$srcdir/$subdir:$objdir:$objdir/../libgcj-3.1.jar"
+ verbose "CLASSPATH is .:$srcdir/$subdir:$objdir:$objdir/../libgcj-$gcc_version.jar"
global env
- set env(CLASSPATH) ".:$srcdir/$subdir:$objdir:$objdir/../libgcj-3.1.jar"
+ set env(CLASSPATH) ".:$srcdir/$subdir:$objdir:$objdir/../libgcj-$gcc_version.jar"
if {$mode == "link"} {
global wrapper_file wrap_compile_flags;
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2002-05-17 21:12 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-05-17 0:19 libgcj-$VERSION changes David S. Miller
2002-05-17 12:33 ` Anthony Green
2002-05-17 14:12 ` David S. Miller
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).