* [patch/testsuite] revert return-type of gdb_file_cmd, gdb_load
@ 2004-09-10 1:03 Michael Chastain
2004-09-10 18:11 ` Andrew Cagney
0 siblings, 1 reply; 3+ messages in thread
From: Michael Chastain @ 2004-09-10 1:03 UTC (permalink / raw)
To: gdb-patches; +Cc: manjo, jimb, cagney
I hope this is the end of the mess, but if it's not, just let me know!
This patch reverts the recent change to the return value of gdb_file_cmd.
Instead of using a structured return value, I just stick the secondary
information into a global variable.
Global variable. Bleagh. But "bleagh" beats "oh shit, I broke eight
other implementations of gdb_load in config/*.exp and I can't test all
of them".
Tested on native i686-pc-linux-gnu, gcc 2.95.3 3.3.4 3.4.1,
dwarf-2 and stabs+, with tcl 8.4.6, expect 5.41, dejagnu 1.4.4.
I also tested gdb.gdb/*.exp with a stripped gdb to check for
the proper "untested" results.
I am committing this now.
===
2004-09-09 Michael Chastain <mec.gnu@mindspring.com>
* lib/gdb.exp (gdb_file_cmd): Revert the return value to
previous simple convention. Use a global variable to store
information about what was found.
(gdb_run_cmd) Adapt to reverted return value.
* gdb.base/remote.exp: Adapt to reverted return value.
* gdb.gdb/complaints.exp: Likewise.
* gdb.gdb/observer.exp: Likewise.
* gdb.gdb/selftest.exp: Likewise.
* gdb.gdb/xfullpath.exp: Likewise.
Index: gdb.base/remote.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/remote.exp,v
retrieving revision 1.4
diff -c -3 -p -r1.4 remote.exp
*** gdb.base/remote.exp 28 Aug 2004 23:55:48 -0000 1.4
--- gdb.base/remote.exp 9 Sep 2004 21:25:13 -0000
*************** proc gdb_load_timed {executable download
*** 119,125 ****
set load_begin_time [clock clicks]
set result [gdb_load $executable]
set load_end_time [clock clicks]
! if { [lindex $result 0] != "" } then {
fail "$test - loading executable"
return
}
--- 119,125 ----
set load_begin_time [clock clicks]
set result [gdb_load $executable]
set load_end_time [clock clicks]
! if { $result != 0 } then {
fail "$test - loading executable"
return
}
Index: gdb.gdb/complaints.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.gdb/complaints.exp,v
retrieving revision 1.4
diff -c -3 -p -r1.4 complaints.exp
*** gdb.gdb/complaints.exp 1 Sep 2004 19:55:00 -0000 1.4
--- gdb.gdb/complaints.exp 9 Sep 2004 21:25:13 -0000
*************** proc setup_test { executable } {
*** 49,65 ****
set timeout 600
verbose "Timeout is now $timeout seconds" 2
set result [gdb_load $executable]
set timeout $oldtimeout
verbose "Timeout is now $timeout seconds" 2
! if { [lindex $result 0] != "" } then {
return -1
- } else {
- if { [lsearch -exact [lrange $result 1 end] "nodebug"] >= 0 } then {
- untested "No debug information, skipping testcase."
- return -1
- }
}
# Set a breakpoint at main
--- 49,68 ----
set timeout 600
verbose "Timeout is now $timeout seconds" 2
+ global gdb_file_cmd_debug_info
+ set gdb_file_cmd_debug_info "unset"
+
set result [gdb_load $executable]
set timeout $oldtimeout
verbose "Timeout is now $timeout seconds" 2
! if { $result != 0 } then {
! return -1
! }
!
! if { gdb_file_cmd_debug_info != "debug" } then {
! untested "No debug information, skipping testcase."
return -1
}
# Set a breakpoint at main
Index: gdb.gdb/observer.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.gdb/observer.exp,v
retrieving revision 1.6
diff -c -3 -p -r1.6 observer.exp
*** gdb.gdb/observer.exp 1 Sep 2004 19:55:00 -0000 1.6
--- gdb.gdb/observer.exp 9 Sep 2004 21:25:13 -0000
*************** proc setup_test { executable } {
*** 47,63 ****
set timeout 600
verbose "Timeout is now $timeout seconds" 2
set result [gdb_load $executable]
set timeout $oldtimeout
verbose "Timeout is now $timeout seconds" 2
! if { [lindex $result 0] != "" } then {
return -1
- } else {
- if { [lsearch -exact [lrange $result 1 end] "nodebug"] >= 0 } then {
- untested "No debug information, skipping testcase."
- return -1
- }
}
# Set a breakpoint at main
--- 47,66 ----
set timeout 600
verbose "Timeout is now $timeout seconds" 2
+ global gdb_file_cmd_debug_info
+ set gdb_file_cmd_debug_info "unset"
+
set result [gdb_load $executable]
set timeout $oldtimeout
verbose "Timeout is now $timeout seconds" 2
! if { $result != 0 } then {
! return -1
! }
!
! if { gdb_file_cmd_debug_info != "debug" } then {
! untested "No debug information, skipping testcase."
return -1
}
# Set a breakpoint at main
Index: gdb.gdb/selftest.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.gdb/selftest.exp,v
retrieving revision 1.5
diff -c -3 -p -r1.5 selftest.exp
*** gdb.gdb/selftest.exp 1 Sep 2004 19:55:00 -0000 1.5
--- gdb.gdb/selftest.exp 9 Sep 2004 21:25:13 -0000
*************** proc test_with_self { executable } {
*** 241,257 ****
set timeout 600
verbose "Timeout is now $timeout seconds" 2
set result [gdb_load $executable]
set timeout $oldtimeout
verbose "Timeout is now $timeout seconds" 2
! if { [lindex $result 0] != "" } then {
return -1
- } else {
- if { [lsearch -exact [lrange $result 1 end] "nodebug"] >= 0 } then {
- untested "No debug information, skipping testcase."
- return -1
- }
}
# disassemble yourself
--- 241,260 ----
set timeout 600
verbose "Timeout is now $timeout seconds" 2
+ global gdb_file_cmd_debug_info
+ set gdb_file_cmd_debug_info "unset"
+
set result [gdb_load $executable]
set timeout $oldtimeout
verbose "Timeout is now $timeout seconds" 2
! if { $result != 0 } then {
! return -1
! }
!
! if { gdb_file_cmd_debug_info != "debug" } then {
! untested "No debug information, skipping testcase."
return -1
}
# disassemble yourself
Index: gdb.gdb/xfullpath.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.gdb/xfullpath.exp,v
retrieving revision 1.4
diff -c -3 -p -r1.4 xfullpath.exp
*** gdb.gdb/xfullpath.exp 1 Sep 2004 19:55:00 -0000 1.4
--- gdb.gdb/xfullpath.exp 9 Sep 2004 21:25:13 -0000
*************** proc setup_test { executable } {
*** 48,64 ****
set timeout 600
verbose "Timeout is now $timeout seconds" 2
set result [gdb_load $executable]
set timeout $oldtimeout
verbose "Timeout is now $timeout seconds" 2
! if { [lindex $result 0] != "" } then {
return -1
- } else {
- if { [lsearch -exact [lrange $result 1 end] "nodebug"] >= 0 } then {
- untested "No debug information, skipping testcase."
- return -1
- }
}
# Set a breakpoint at main
--- 48,67 ----
set timeout 600
verbose "Timeout is now $timeout seconds" 2
+ global gdb_file_cmd_debug_info
+ set gdb_file_cmd_debug_info "unset"
+
set result [gdb_load $executable]
set timeout $oldtimeout
verbose "Timeout is now $timeout seconds" 2
! if { $result != 0 } then {
! return -1
! }
!
! if { gdb_file_cmd_debug_info != "debug" } then {
! untested "No debug information, skipping testcase."
return -1
}
# Set a breakpoint at main
Index: lib/gdb.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/lib/gdb.exp,v
retrieving revision 1.57
diff -c -3 -p -r1.57 gdb.exp
*** lib/gdb.exp 28 Aug 2004 23:55:48 -0000 1.57
--- lib/gdb.exp 9 Sep 2004 21:25:14 -0000
*************** proc gdb_run_cmd {args} {
*** 180,187 ****
if [target_info exists gdb,do_reload_on_run] {
# Specifying no file, defaults to the executable
# currently being debugged.
! set status [gdb_load ""]
! if { [lindex $status 0] != "" } {
return;
}
send_gdb "continue\n";
--- 180,186 ----
if [target_info exists gdb,do_reload_on_run] {
# Specifying no file, defaults to the executable
# currently being debugged.
! if { [gdb_load ""] != 0 } {
return;
}
send_gdb "continue\n";
*************** proc gdb_run_cmd {args} {
*** 226,233 ****
send_gdb "y\n"
}
-re "The program is not being run.*$gdb_prompt $" {
! set status [gdb_load ""]
! if { [lindex $status 0] != ""] } {
return;
}
send_gdb "jump *$start\n";
--- 225,231 ----
send_gdb "y\n"
}
-re "The program is not being run.*$gdb_prompt $" {
! if { [gdb_load ""] != 0 } {
return;
}
send_gdb "jump *$start\n";
*************** proc gdb_run_cmd {args} {
*** 249,256 ****
}
if [target_info exists gdb,do_reload_on_run] {
! set status [gdb_load ""]
! if { [lindex $status 0] != "" } {
return;
}
}
--- 247,253 ----
}
if [target_info exists gdb,do_reload_on_run] {
! if { [gdb_load ""] != 0 } {
return;
}
}
*************** proc default_gdb_exit {} {
*** 957,992 ****
}
# Load a file into the debugger.
! # The return value is a list with the following information:
#
! # { message word ... }
! #
! # MESSAGE has the following values:
! #
! # "" file was loaded successfully
! # "..." file was not loaded successfully.
! # A perror has been generated with MESSAGE.
#
! # If the MESSAGE is "", then there is an optional set of words.
! # The words may be:
#
! # nodebug this file does not contain debug information
#
! # TODO: gdb.base/sepdebug.exp and gdb.stabs/weird.exp might
! # be able to use this if they can get more information
! # in the return value.
proc gdb_file_cmd { arg } {
global gdb_prompt
global verbose
global GDB
if [is_remote host] {
set arg [remote_download host $arg]
if { $arg == "" } {
! set message "download failed"
! perror $message
! return { $message }
}
}
--- 954,990 ----
}
# Load a file into the debugger.
! # The return value is 0 for success, -1 for failure.
#
! # This procedure also set the global variable GDB_FILE_CMD_DEBUG_INFO
! # to one of these values:
#
! # debug file was loaded successfully and has debug information
! # nodebug file was loaded successfully and has no debug information
! # fail file was not loaded
#
! # I tried returning this information as part of the return value,
! # but ran into a mess because of the many re-implementations of
! # gdb_load in config/*.exp.
#
! # TODO: gdb.base/sepdebug.exp and gdb.stabs/weird.exp might be able to use
! # this if they can get more information set.
proc gdb_file_cmd { arg } {
global gdb_prompt
global verbose
global GDB
+ # Set whether debug info was found.
+ # Default to "fail".
+ global gdb_file_cmd_debug_info
+ set gdb_file_cmd_debug_info "fail"
+
if [is_remote host] {
set arg [remote_download host $arg]
if { $arg == "" } {
! perror "download failed"
! return -1
}
}
*************** proc gdb_file_cmd { arg } {
*** 994,1008 ****
gdb_expect 120 {
-re "Reading symbols from.*no debugging symbols found.*done.*$gdb_prompt $" {
verbose "\t\tLoaded $arg into the $GDB with no debugging symbols"
! return { "" nodebug }
}
-re "Reading symbols from.*done.*$gdb_prompt $" {
verbose "\t\tLoaded $arg into the $GDB"
! return { "" }
}
-re "A program is being debugged already.*Kill it.*y or n. $" {
send_gdb "y\n"
! verbose "\t\tKilling previous program being debugged"
exp_continue
}
-re "Load new symbol table from \".*\".*y or n. $" {
--- 992,1008 ----
gdb_expect 120 {
-re "Reading symbols from.*no debugging symbols found.*done.*$gdb_prompt $" {
verbose "\t\tLoaded $arg into the $GDB with no debugging symbols"
! set gdb_file_cmd_debug_info "nodebug"
! return 0
}
-re "Reading symbols from.*done.*$gdb_prompt $" {
verbose "\t\tLoaded $arg into the $GDB"
! set gdb_file_cmd_debug_info "debug"
! return 0
}
-re "A program is being debugged already.*Kill it.*y or n. $" {
send_gdb "y\n"
! verbose "\t\tKilling previous program being debugged"
exp_continue
}
-re "Load new symbol table from \".*\".*y or n. $" {
*************** proc gdb_file_cmd { arg } {
*** 1010,1046 ****
gdb_expect 120 {
-re "Reading symbols from.*done.*$gdb_prompt $" {
verbose "\t\tLoaded $arg with new symbol table into $GDB"
! return { "" }
}
timeout {
! set message "(timeout) Couldn't load $arg, other program already loaded."
! perror $message
! return { $message }
}
}
}
-re "No such file or directory.*$gdb_prompt $" {
! set message "($arg) No such file or directory"
! perror $message
! return { $message }
}
-re "$gdb_prompt $" {
! set message "couldn't load $arg into $GDB."
! perror $message
! return { $message }
}
timeout {
! set message "couldn't load $arg into $GDB (timed out)."
! perror $message
! return { $message }
}
eof {
# This is an attempt to detect a core dump, but seems not to
# work. Perhaps we need to match .* followed by eof, in which
# gdb_expect does not seem to have a way to do that.
! set message "couldn't load $arg into $GDB (end of file)."
! perror $message
! return { $message }
}
}
}
--- 1010,1042 ----
gdb_expect 120 {
-re "Reading symbols from.*done.*$gdb_prompt $" {
verbose "\t\tLoaded $arg with new symbol table into $GDB"
! set gdb_file_cmd_debug_info "debug"
! return 0
}
timeout {
! perror "(timeout) Couldn't load $arg, other program already loaded."
! return -1
}
}
}
-re "No such file or directory.*$gdb_prompt $" {
! perror "($arg) No such file or directory"
! return -1
}
-re "$gdb_prompt $" {
! perror "couldn't load $arg into $GDB."
! return -1
}
timeout {
! perror "couldn't load $arg into $GDB (timed out)."
! return -1
}
eof {
# This is an attempt to detect a core dump, but seems not to
# work. Perhaps we need to match .* followed by eof, in which
# gdb_expect does not seem to have a way to do that.
! perror "couldn't load $arg into $GDB (end of file)."
! return -1
}
}
}
*************** proc gdb_exit { } {
*** 1655,1660 ****
--- 1651,1657 ----
#
# gdb_load -- load a file into the debugger.
+ # Many files in config/*.exp override this procedure.
#
proc gdb_load { arg } {
return [gdb_file_cmd $arg]
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [patch/testsuite] revert return-type of gdb_file_cmd, gdb_load
2004-09-10 1:03 [patch/testsuite] revert return-type of gdb_file_cmd, gdb_load Michael Chastain
@ 2004-09-10 18:11 ` Andrew Cagney
2004-09-10 18:30 ` Michael Chastain
0 siblings, 1 reply; 3+ messages in thread
From: Andrew Cagney @ 2004-09-10 18:11 UTC (permalink / raw)
To: Michael Chastain; +Cc: gdb-patches
> I hope this is the end of the mess, but if it's not, just let me know!
Thanks! (Joel, my mips patches will no longer be flying blind ;-)
> This patch reverts the recent change to the return value of gdb_file_cmd.
> Instead of using a structured return value, I just stick the secondary
> information into a global variable.
What of two interfaces -> an existing one with preserved behavior and a
new one with extended behavior?
> Global variable. Bleagh. But "bleagh" beats "oh shit, I broke eight
> other implementations of gdb_load in config/*.exp and I can't test all
> of them".
>
> Tested on native i686-pc-linux-gnu, gcc 2.95.3 3.3.4 3.4.1,
> dwarf-2 and stabs+, with tcl 8.4.6, expect 5.41, dejagnu 1.4.4.
>
> I also tested gdb.gdb/*.exp with a stripped gdb to check for
> the proper "untested" results.
>
> I am committing this now.
Andrew
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [patch/testsuite] revert return-type of gdb_file_cmd, gdb_load
2004-09-10 18:11 ` Andrew Cagney
@ 2004-09-10 18:30 ` Michael Chastain
0 siblings, 0 replies; 3+ messages in thread
From: Michael Chastain @ 2004-09-10 18:30 UTC (permalink / raw)
To: cagney; +Cc: gdb-patches
Andrew Cagney <cagney@gnu.org> wrote:
> What of two interfaces -> an existing one with preserved behavior and a
> new one with extended behavior?
In a way, that's what we have now. The "extended behavior" is to set
the global variable gdb_file_cmd_debug_info.
The problem with "two interfaces" is that there really are nine
different procedures named gdb_load, and the caller does not know
which of the nine procedures is active. So the new interface has
to be compatible with the old, or I have to change nine implementations
of "gdb_load" simultaneously.
Also, if gdb.gdb/*.exp runs in an environment where the implementation of
gdb_load does not provide this extended behavior, then it will
see that gdb_file_cmd_debug_info is still set to "unset" and it can
react appropriately. From inspection, I believe the only gdb_load
that will *not* get the extended behavior is the stubby
"not implemented yet" implementation in config/nind.exp.
Michael
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2004-09-10 18:30 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-09-10 1:03 [patch/testsuite] revert return-type of gdb_file_cmd, gdb_load Michael Chastain
2004-09-10 18:11 ` Andrew Cagney
2004-09-10 18:30 ` Michael Chastain
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox