[1/5,gdb/testsuite] Fix some test-cases for check-read1 (gdb_test_lines)
Checks
| Context |
Check |
Description |
| linaro-tcwg-bot/tcwg_gdb_build--master-aarch64 |
success
|
Build passed
|
| linaro-tcwg-bot/tcwg_gdb_build--master-arm |
success
|
Build passed
|
| linaro-tcwg-bot/tcwg_gdb_check--master-arm |
success
|
Test passed
|
| linaro-tcwg-bot/tcwg_gdb_check--master-aarch64 |
success
|
Test passed
|
Commit Message
I ran the testsuite in an environment simulating a stressed system in
combination with check-read1. This exposes a few more FAILs.
Fix some by using gdb_test_lines, as well as related gdb_get_lines.
Tested on x86_64-linux.
---
.../premature-dummy-frame-removal.exp | 14 ++--
gdb/testsuite/gdb.btrace/exception.exp | 28 ++++----
gdb/testsuite/gdb.cp/overload.exp | 69 +++++++++----------
gdb/testsuite/gdb.python/py-framefilter.exp | 22 +++---
gdb/testsuite/gdb.reverse/sigall-precsave.exp | 4 +-
.../gdb.server/bkpt-other-inferior.exp | 30 ++++----
gdb/testsuite/gdb.server/ext-run.exp | 4 +-
7 files changed, 83 insertions(+), 88 deletions(-)
base-commit: 1001055e3552760015661dd9b081435051fe1340
Comments
Hello Tom,
>diff --git a/gdb/testsuite/gdb.btrace/exception.exp
>b/gdb/testsuite/gdb.btrace/exception.exp
>index 61cc6c904b6..f27e40f2bcd 100755
>--- a/gdb/testsuite/gdb.btrace/exception.exp
>+++ b/gdb/testsuite/gdb.btrace/exception.exp
>@@ -64,17 +64,17 @@ gdb_expect_list "flat" "\r\n$gdb_prompt $" [list \
> ]
>
> # show the branch trace with calls indented
>-send_gdb "record function-call-history /c 1\n"
>-gdb_expect_list "indented" "\r\n$gdb_prompt $" [list \
>- [multi_line \
>- "1\tmain\\(\\)" \
>- "2\t test\\(\\)" \
>- "3\t foo\\(\\)" \
>- "4\t bar\\(\\)" \
>- "5\t bad\\(\\)\r" \
>- ] "" \
>- [multi_line \
>- "\[0-9\]*\t test\\(\\)" \
>- "\[0-9\]*\tmain\\(\\)" \
>- ] "" \
>- ]
>+set lines [gdb_get_lines "record function-call-history /c 1"]
>+set re1 \
>+ [multi_line \
>+ "1\tmain\\(\\)" \
>+ "2\t test\\(\\)" \
>+ "3\t foo\\(\\)" \
>+ "4\t bar\\(\\)" \
>+ "5\t bad\\(\\)\r"]
>+set re2 \
>+ [multi_line \
>+ "\[0-9\]*\t test\\(\\)" \
>+ "\[0-9\]*\tmain\\(\\)"]
>+
>+gdb_assert { [regexp $re1.*$re2 $lines] } "indented"
Could this be handled by gdb_test_sequence instead?
If not, should another gdb_test_<something> be introduced to implement
the testing pattern you applied here?
Regards,
Markus.
Intel Deutschland GmbH
Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Sean Fennelly, Jeffrey Schneiderman, Tiffany Doon Silva
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928
On 10/10/24 09:35, Metzger, Markus T wrote:
> Hello Tom,
>
>> diff --git a/gdb/testsuite/gdb.btrace/exception.exp
>> b/gdb/testsuite/gdb.btrace/exception.exp
>> index 61cc6c904b6..f27e40f2bcd 100755
>> --- a/gdb/testsuite/gdb.btrace/exception.exp
>> +++ b/gdb/testsuite/gdb.btrace/exception.exp
>> @@ -64,17 +64,17 @@ gdb_expect_list "flat" "\r\n$gdb_prompt $" [list \
>> ]
>>
>> # show the branch trace with calls indented
>> -send_gdb "record function-call-history /c 1\n"
>> -gdb_expect_list "indented" "\r\n$gdb_prompt $" [list \
>> - [multi_line \
>> - "1\tmain\\(\\)" \
>> - "2\t test\\(\\)" \
>> - "3\t foo\\(\\)" \
>> - "4\t bar\\(\\)" \
>> - "5\t bad\\(\\)\r" \
>> - ] "" \
>> - [multi_line \
>> - "\[0-9\]*\t test\\(\\)" \
>> - "\[0-9\]*\tmain\\(\\)" \
>> - ] "" \
>> - ]
>> +set lines [gdb_get_lines "record function-call-history /c 1"]
>> +set re1 \
>> + [multi_line \
>> + "1\tmain\\(\\)" \
>> + "2\t test\\(\\)" \
>> + "3\t foo\\(\\)" \
>> + "4\t bar\\(\\)" \
>> + "5\t bad\\(\\)\r"]
>> +set re2 \
>> + [multi_line \
>> + "\[0-9\]*\t test\\(\\)" \
>> + "\[0-9\]*\tmain\\(\\)"]
>> +
>> +gdb_assert { [regexp $re1.*$re2 $lines] } "indented"
>
> Could this be handled by gdb_test_sequence instead?
>
Hi Markus,
thanks for the review.
Proc gdb_test_sequence is a wrapper around gdb_expect_list, so it has
the same problem.
I avoid using gdb_test_sequence because it does not work well with large
outputs (as demonstrated by this test-case, and other test-cases in the
same patch). It does solve part of the problem: rather that matching
^$re1.*$re2 in one action, it splits this up in two actions, matching
^$re1 and .*$re2, but the .* part can still be large and produce timeouts.
> If not, should another gdb_test_<something> be introduced to implement
> the testing pattern you applied here?
Um, I'm not entirely sure this is what you mean, but gdb_get_lines was
factored out of gdb_test_lines, and in this case this could be used
instead, giving a more regular:
...
gdb_test_lines "record function-call-history /c 1" \
"indented" $re1.*$re2
...
I'll fix this in my local branch.
Thanks,
- Tom
Hello Tom,
>>> # show the branch trace with calls indented
>>> -send_gdb "record function-call-history /c 1\n"
>>> -gdb_expect_list "indented" "\r\n$gdb_prompt $" [list \
>>> - [multi_line \
>>> - "1\tmain\\(\\)" \
>>> - "2\t test\\(\\)" \
>>> - "3\t foo\\(\\)" \
>>> - "4\t bar\\(\\)" \
>>> - "5\t bad\\(\\)\r" \
>>> - ] "" \
>>> - [multi_line \
>>> - "\[0-9\]*\t test\\(\\)" \
>>> - "\[0-9\]*\tmain\\(\\)" \
>>> - ] "" \
>>> - ]
>>> +set lines [gdb_get_lines "record function-call-history /c 1"]
>>> +set re1 \
>>> + [multi_line \
>>> + "1\tmain\\(\\)" \
>>> + "2\t test\\(\\)" \
>>> + "3\t foo\\(\\)" \
>>> + "4\t bar\\(\\)" \
>>> + "5\t bad\\(\\)\r"]
>>> +set re2 \
>>> + [multi_line \
>>> + "\[0-9\]*\t test\\(\\)" \
>>> + "\[0-9\]*\tmain\\(\\)"]
>>> +
>>> +gdb_assert { [regexp $re1.*$re2 $lines] } "indented"
>>
>> Could this be handled by gdb_test_sequence instead?
>>
>Proc gdb_test_sequence is a wrapper around gdb_expect_list, so it has
>the same problem.
>
>I avoid using gdb_test_sequence because it does not work well with large
>outputs (as demonstrated by this test-case, and other test-cases in the
>same patch). It does solve part of the problem: rather that matching
>^$re1.*$re2 in one action, it splits this up in two actions, matching
>^$re1 and .*$re2, but the .* part can still be large and produce timeouts.
We're not really interested in the .* part so it's a pity we're storing it.
IIRC I ended up using gdb_test_multiple in other contexts where I wanted
to discard some huge irrelevant output in the middle that caused sporadic
buffer overflows. That output was uniform so I could write an extra -re
case for it.
IIUC gdb_expect_list tries to match one pattern at a time. Could we add a
generic -re at the end of the gdb_expect that matches and discards one line?
Since the multi-line pattern we want to match appears first, I hope it would
also match first and we don't end up with the discard pattern discarding one
line of the output we want to match at a time so that the multi-line pattern
doesn't get a chance to match.
>> If not, should another gdb_test_<something> be introduced to implement
>> the testing pattern you applied here?
>
>Um, I'm not entirely sure this is what you mean, but gdb_get_lines was
>factored out of gdb_test_lines, and in this case this could be used
>instead, giving a more regular:
>...
>gdb_test_lines "record function-call-history /c 1" \
> "indented" $re1.*$re2
Would that allow us to keep the original structure of the test?
Thanks,
Markus.
Intel Deutschland GmbH
Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Sean Fennelly, Jeffrey Schneiderman, Tiffany Doon Silva
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928
On 10/10/24 10:27, Metzger, Markus T wrote:
> Hello Tom,
>
>>>> # show the branch trace with calls indented
>>>> -send_gdb "record function-call-history /c 1\n"
>>>> -gdb_expect_list "indented" "\r\n$gdb_prompt $" [list \
>>>> - [multi_line \
>>>> - "1\tmain\\(\\)" \
>>>> - "2\t test\\(\\)" \
>>>> - "3\t foo\\(\\)" \
>>>> - "4\t bar\\(\\)" \
>>>> - "5\t bad\\(\\)\r" \
>>>> - ] "" \
>>>> - [multi_line \
>>>> - "\[0-9\]*\t test\\(\\)" \
>>>> - "\[0-9\]*\tmain\\(\\)" \
>>>> - ] "" \
>>>> - ]
>>>> +set lines [gdb_get_lines "record function-call-history /c 1"]
>>>> +set re1 \
>>>> + [multi_line \
>>>> + "1\tmain\\(\\)" \
>>>> + "2\t test\\(\\)" \
>>>> + "3\t foo\\(\\)" \
>>>> + "4\t bar\\(\\)" \
>>>> + "5\t bad\\(\\)\r"]
>>>> +set re2 \
>>>> + [multi_line \
>>>> + "\[0-9\]*\t test\\(\\)" \
>>>> + "\[0-9\]*\tmain\\(\\)"]
>>>> +
>>>> +gdb_assert { [regexp $re1.*$re2 $lines] } "indented"
>>>
>>> Could this be handled by gdb_test_sequence instead?
>>>
>> Proc gdb_test_sequence is a wrapper around gdb_expect_list, so it has
>> the same problem.
>>
>> I avoid using gdb_test_sequence because it does not work well with large
>> outputs (as demonstrated by this test-case, and other test-cases in the
>> same patch). It does solve part of the problem: rather that matching
>> ^$re1.*$re2 in one action, it splits this up in two actions, matching
>> ^$re1 and .*$re2, but the .* part can still be large and produce timeouts.
>
> We're not really interested in the .* part so it's a pity we're storing it.
>
True, though from my experience of writing this series, in some cases
that seems to be factors faster than leaving it in the matching buffer.
> IIRC I ended up using gdb_test_multiple in other contexts where I wanted
> to discard some huge irrelevant output in the middle that caused sporadic
> buffer overflows. That output was uniform so I could write an extra -re
> case for it.
>
> IIUC gdb_expect_list tries to match one pattern at a time. Could we add a
> generic -re at the end of the gdb_expect that matches and discards one line?
>
> Since the multi-line pattern we want to match appears first, I hope it would
> also match first and we don't end up with the discard pattern discarding one
> line of the output we want to match at a time so that the multi-line pattern
> doesn't get a chance to match.
>
Well, gdb_test and gdb_test_multiple have functionality to match
line-by-line, using -lbl. But that means multi-line patterns (like say
-re -wrap "bla") don't work reliably.
Sometimes we can work around this by massaging the regexps.
In the case of patch "[gdb/testsuite] Fix gdb.threads/ia64-sigill.exp
with check-read1" from this series, I couldn't so I wrote a custom
line-by-line clause:
...
-re "\r\n\[^\r\n\]*(?=\r\n\[^\r\n\]*\r\n)" {
exp_continue
}
...
that still allows matching two lines.
>>> If not, should another gdb_test_<something> be introduced to implement
>>> the testing pattern you applied here?
>>
>> Um, I'm not entirely sure this is what you mean, but gdb_get_lines was
>> factored out of gdb_test_lines, and in this case this could be used
>> instead, giving a more regular:
>> ...
>> gdb_test_lines "record function-call-history /c 1" \
>> "indented" $re1.*$re2
>
> Would that allow us to keep the original structure of the test?
Sorry, I don't understand what you mean. What do you consider the
original structure that's no longer there? What is your concern?
Thanks,
- Tom
>>> Um, I'm not entirely sure this is what you mean, but gdb_get_lines was
>>> factored out of gdb_test_lines, and in this case this could be used
>>> instead, giving a more regular:
>>> ...
>>> gdb_test_lines "record function-call-history /c 1" \
>>> "indented" $re1.*$re2
>>
>> Would that allow us to keep the original structure of the test?
>
>Sorry, I don't understand what you mean. What do you consider the
>original structure that's no longer there? What is your concern?
The original code used
gdb_expect_list "indented" "\r\n$gdb_prompt $" [list \
[multi_line \
"1\tmain\\(\\)" \
"2\t test\\(\\)" \
"3\t foo\\(\\)" \
"4\t bar\\(\\)" \
"5\t bad\\(\\)\r" \
] "" \
[multi_line \
"\[0-9\]*\t test\\(\\)" \
"\[0-9\]*\tmain\\(\\)" \
] "" \
]
I was wondering if we could turn this into
gdb_test_lines "record function-call-history /c 1" \
"indented" \
[multi_line \
"1\tmain\\(\\)" \
"2\t test\\(\\)" \
"3\t foo\\(\\)" \
"4\t bar\\(\\)" \
"5\t bad\\(\\)\r" \
] \
.*\
[multi_line \
"\[0-9\]*\t test\\(\\)" \
"\[0-9\]*\tmain\\(\\)" \
]
Regards,
Markus.
Intel Deutschland GmbH
Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Sean Fennelly, Jeffrey Schneiderman, Tiffany Doon Silva
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928
On 10/10/24 11:03, Metzger, Markus T wrote:
> I was wondering if we could turn this into
>
> gdb_test_lines "record function-call-history /c 1" \
> "indented" \
> [multi_line \
> "1\tmain\\(\\)" \
> "2\t test\\(\\)" \
> "3\t foo\\(\\)" \
> "4\t bar\\(\\)" \
> "5\t bad\\(\\)\r" \
> ] \
> .*\
> [multi_line \
> "\[0-9\]*\t test\\(\\)" \
> "\[0-9\]*\tmain\\(\\)" \
> ]
Proc gdb_test_lines only takes a single regexp argument, and this passes
3 of them.
I can fix that by doing (btw also using $decimal as additional cleanup):
...
gdb_test_lines "record function-call-history /c 1" \
"indented" \
[multi_line \
"1\tmain\\(\\)" \
"2\t test\\(\\)" \
"3\t foo\\(\\)" \
"4\t bar\\(\\)" \
"5\t bad\\(\\)" \
".*" \
"$decimal\t test\\(\\)" \
"$decimal\tmain\\(\\)"]
...
Thanks,
- Tom
>> I was wondering if we could turn this into
>>
>> gdb_test_lines "record function-call-history /c 1" \
>> "indented" \
>> [multi_line \
>> "1\tmain\\(\\)" \
>> "2\t test\\(\\)" \
>> "3\t foo\\(\\)" \
>> "4\t bar\\(\\)" \
>> "5\t bad\\(\\)\r" \
>> ] \
>> .*\
>> [multi_line \
>> "\[0-9\]*\t test\\(\\)" \
>> "\[0-9\]*\tmain\\(\\)" \
>> ]
>
>
>Proc gdb_test_lines only takes a single regexp argument, and this passes
>3 of them.
>
>I can fix that by doing (btw also using $decimal as additional cleanup):
>...
>gdb_test_lines "record function-call-history /c 1" \
> "indented" \
> [multi_line \
> "1\tmain\\(\\)" \
> "2\t test\\(\\)" \
> "3\t foo\\(\\)" \
> "4\t bar\\(\\)" \
> "5\t bad\\(\\)" \
> ".*" \
> "$decimal\t test\\(\\)" \
> "$decimal\tmain\\(\\)"]
>...
Thanks, that looks good.
Markus.
Intel Deutschland GmbH
Registered Address: Am Campeon 10, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Sean Fennelly, Jeffrey Schneiderman, Tiffany Doon Silva
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928
On 10/10/24 09:25, Tom de Vries wrote:
> I ran the testsuite in an environment simulating a stressed system in
> combination with check-read1. This exposes a few more FAILs.
>
> Fix some by using gdb_test_lines, as well as related gdb_get_lines.
>
I've pushed this series.
Thanks,
- Tom
@@ -80,12 +80,10 @@ gdb_test_multiple "p some_func ()" "" {
# debug, to format of which isn't fixed. All we care about is that
# GDB is still running afterwards.
#
-# All of the debug output makes this really slow when testing with the
-# special read1 version of expect, hence the timeout factor.
-with_read1_timeout_factor 10 {
- gdb_test_no_output "set debug frame on"
- gdb_test "p some_func ()" ".*" \
- "repeat p some_func () with frame debug on"
- gdb_test_no_output "set debug frame off"
-}
+gdb_test_no_output "set debug frame on"
+gdb_test_lines "p some_func ()" \
+ "repeat p some_func () with frame debug on" \
+ ".*"
+gdb_test_no_output "set debug frame off"
+
gdb_test "p 1 + 2 + 3" " = 6"
@@ -64,17 +64,17 @@ gdb_expect_list "flat" "\r\n$gdb_prompt $" [list \
]
# show the branch trace with calls indented
-send_gdb "record function-call-history /c 1\n"
-gdb_expect_list "indented" "\r\n$gdb_prompt $" [list \
- [multi_line \
- "1\tmain\\(\\)" \
- "2\t test\\(\\)" \
- "3\t foo\\(\\)" \
- "4\t bar\\(\\)" \
- "5\t bad\\(\\)\r" \
- ] "" \
- [multi_line \
- "\[0-9\]*\t test\\(\\)" \
- "\[0-9\]*\tmain\\(\\)" \
- ] "" \
- ]
+set lines [gdb_get_lines "record function-call-history /c 1"]
+set re1 \
+ [multi_line \
+ "1\tmain\\(\\)" \
+ "2\t test\\(\\)" \
+ "3\t foo\\(\\)" \
+ "4\t bar\\(\\)" \
+ "5\t bad\\(\\)\r"]
+set re2 \
+ [multi_line \
+ "\[0-9\]*\t test\\(\\)" \
+ "\[0-9\]*\tmain\\(\\)"]
+
+gdb_assert { [regexp $re1.*$re2 $lines] } "indented"
@@ -89,42 +89,39 @@ set re_synth "foo & operator=\\(foo const ?&\\);"
gdb_test "print foo_instance1" "\\$\[0-9\]+ = \{ifoo = 111, ccpfoo = 0x0\}"
-gdb_test_multiple "ptype foo_instance1" "ptype foo_instance1" {
- -re "type = $re_class${ws}$XX_fields${ws}$re_synth${ws}$re_dtor${ws}$re_ctor${ws}$re_methods$nl\}$nl$gdb_prompt $" {
- # gcc 2.95.3 -gstabs+, no "const" on "const char *"
- # TODO: gdb.base/constvar.exp has XFAILed this kind of problem for a
- # long time, but an XFAIL really needs an external bug report.
- # -- chastain 2003-12-31
- # setup_xfail "*-*-*"
- # fail "ptype foo_instance1"
- # TODO: this should be a KFAIL.
- pass "ptype foo_instance1 (shorter match)"
- }
- -re "type = $re_class${ws}$re_fields${ws}$re_synth${ws}$re_dtor${ws}$re_ctor${ws}$re_methods$nl\}$nl$gdb_prompt $" {
- # gcc 2.95.3 -gstabs+ if "const char *" ever gets fixed
- pass "ptype foo_instance1"
- }
- -re "type = $re_class${ws}$re_fields${ws}$re_ctor${ws}$XX_dtor${ws}$re_methods$nl\}$nl$gdb_prompt $" {
- # gcc 3.3.2 -gdwarf-2, "~foo(int)"
- # TODO: kfail this
- # kfail "gdb/1113" "ptype foo_instance1"
- pass "ptype foo_instance1 (shorter match)"
- }
- -re "type = $re_class${ws}$re_fields${ws}$re_ctor${ws}$re_dtor${ws}$re_methods$nl\}$nl$gdb_prompt $" {
- # gcc 3.3.2 -gdwarf-2, if the dtor bug gets fixed
- # gcc HEAD -gdwarf-2 (abi-2)
- # TODO: just pass this
- pass "ptype foo_instance1 (shorter match)"
- }
- -re "type = $re_class${ws}$re_fields${ws}$re_synth${ws}$re_ctor${ws}$re_dtor${ws}$re_methods$nl\}$nl$gdb_prompt $" {
- # gcc 3.3.2 -gstabs+
- # TODO: enough with the "shorter match"
- pass "ptype foo_instance1 (shorter match)"
- }
- -re "type = $re_class${ws}$re_fields${ws}$re_ctor${ws}$re_dtor${ws}$re_methods${ws}$re_synth$nl\}$nl$gdb_prompt $" {
- # gcc HEAD -gstabs+ (abi-2)
- pass "ptype foo_instance1 (shorter match)"
- }
+set lines [gdb_get_lines "ptype foo_instance1"]
+set test "ptype foo_instance1 output"
+if { [regexp "^type = $re_class${ws}$XX_fields${ws}$re_synth${ws}$re_dtor${ws}$re_ctor${ws}$re_methods$nl\}\r\n$" $lines] } {
+ # gcc 2.95.3 -gstabs+, no "const" on "const char *"
+ # TODO: gdb.base/constvar.exp has XFAILed this kind of problem for a
+ # long time, but an XFAIL really needs an external bug report.
+ # -- chastain 2003-12-31
+ # setup_xfail "*-*-*"
+ # fail "ptype foo_instance1"
+ # TODO: this should be a KFAIL.
+ pass "$test (shorter match)"
+} elseif { [regexp "^type = $re_class${ws}$re_fields${ws}$re_synth${ws}$re_dtor${ws}$re_ctor${ws}$re_methods$nl\}\r\n$" $lines] } {
+ # gcc 2.95.3 -gstabs+ if "const char *" ever gets fixed
+ pass "$test"
+} elseif { [regexp "^type = $re_class${ws}$re_fields${ws}$re_ctor${ws}$XX_dtor${ws}$re_methods$nl\}\r\n$" $lines] } {
+ # gcc 3.3.2 -gdwarf-2, "~foo(int)"
+ # TODO: kfail this
+ # kfail "gdb/1113" "ptype foo_instance1"
+ pass "$test (shorter match)"
+} elseif { [regexp "^type = $re_class${ws}$re_fields${ws}$re_ctor${ws}$re_dtor${ws}$re_methods$nl\}\r\n$" $lines] } {
+ # gcc 3.3.2 -gdwarf-2, if the dtor bug gets fixed
+ # gcc HEAD -gdwarf-2 (abi-2)
+ # TODO: just pass this
+ pass "$test (shorter match)"
+} elseif { [regexp "^type = $re_class${ws}$re_fields${ws}$re_synth${ws}$re_ctor${ws}$re_dtor${ws}$re_methods$nl\}\r\n$" $lines] } {
+ # gcc 3.3.2 -gstabs+
+ # TODO: enough with the "shorter match"
+ pass "$test (shorter match)"
+} elseif { [regexp "^type = $re_class${ws}$re_fields${ws}$re_ctor${ws}$re_dtor${ws}$re_methods${ws}$re_synth$nl\}\r\n$" $lines] } {
+ # gcc HEAD -gstabs+ (abi-2)
+ pass "$test (shorter match)"
+} else {
+ fail $test
}
# Print variables and method calls.
@@ -82,11 +82,10 @@ gdb_breakpoint [gdb_get_line_number "Inner test breakpoint"]
gdb_continue_to_breakpoint "Inner test breakpoint"
# Test multiple local blocks.
-gdb_test "bt full no-filters" \
+gdb_test_lines "bt full no-filters" "" \
".*#0.*end_func.*h = 9.*f = 42.*g = 19.*bar = $hex \"Inside block x2\".*d = 15.*e = 14.*foo = $hex \"Inside block\".*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*"
-gdb_test "bt full" \
- ".*#0.*cnuf_dne.*h = 9.*f = 42.*g = 19.*bar = $hex \"Inside block x2\".*d = 15.*e = 14.*foo = $hex \"Inside block\".*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*" \
- "bt full with filters"
+gdb_test_lines "bt full" "bt full with filters" \
+ ".*#0.*cnuf_dne.*h = 9.*f = 42.*g = 19.*bar = $hex \"Inside block x2\".*d = 15.*e = 14.*foo = $hex \"Inside block\".*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*"
# Test pagination can be aborted even for frame filters.
gdb_test_no_output "set height 5" "pagination quit - set height limited"
@@ -161,15 +160,12 @@ gdb_test "bt -2" \
gdb_test "bt 3" \
".*#0.*end_func.*#1.*in funca \\(\\).*#2.*in funcb \\(j=10\\)\[^#\]*More stack frames follow.*" \
"bt 3 with frame-filter Reverse disabled"
-gdb_test "bt no-filter full" \
- ".*#0.*end_func.*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*#1.*in funca \\(\\).*#2.*in funcb \\(j=10\\).*bar = \{a = 42, b = 84\}.*" \
- "bt no-filters full with Reverse disabled"
-gdb_test "bt full" \
- ".*#0.*end_func.*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*#1.*in funca \\(\\).*#2.*in funcb \\(j=10\\).*bar = \{a = 42, b = 84\}.*#22.*in func1 \\(\\).*#23.*in func2 \\(f=3\\).*elided = $hex \"Elided frame\".*fb = \{nothing = $hex \"Elided Foo Bar\", f = 84, s = 38\}.*bf = $hex.*" \
- "bt full with Reverse disabled"
-gdb_test "bt full hide" \
- ".*#0.*end_func.*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*#1.*in funca \\(\\).*#2.*in funcb \\(j=10\\).*bar = \{a = 42, b = 84\}.*#22.*in func1 \\(\\)\[^#\]*#24.*in func3 \\(i=3\\).*" \
- "bt full hide with Reverse disabled"
+gdb_test_lines "bt no-filter full" "bt no-filters full with Reverse disabled" \
+ ".*#0.*end_func.*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*#1.*in funca \\(\\).*#2.*in funcb \\(j=10\\).*bar = \{a = 42, b = 84\}.*"
+gdb_test_lines "bt full" "bt full with Reverse disabled" \
+ ".*#0.*end_func.*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*#1.*in funca \\(\\).*#2.*in funcb \\(j=10\\).*bar = \{a = 42, b = 84\}.*#22.*in func1 \\(\\).*#23.*in func2 \\(f=3\\).*elided = $hex \"Elided frame\".*fb = \{nothing = $hex \"Elided Foo Bar\", f = 84, s = 38\}.*bf = $hex.*"
+gdb_test_lines "bt full hide" "bt full hide with Reverse disabled" \
+ ".*#0.*end_func.*str = $hex \"The End\".*st2 = $hex \"Is Near\".*b = 12.*c = 5.*#1.*in funca \\(\\).*#2.*in funcb \\(j=10\\).*bar = \{a = 42, b = 84\}.*#22.*in func1 \\(\\)\[^#\]*#24.*in func3 \\(i=3\\).*"
# Re-enable Reverse
gdb_test_no_output "enable frame-filter global Reverse" \
@@ -155,7 +155,7 @@ gdb_test "break $breakloc" \
"breakpoint at end of main"
# Signal handlers must be disabled
-gdb_test "handle all nostop noprint"
+gdb_test_lines "handle all nostop noprint" "" ""
# The list of signals that the program generates, in the order they
# are generated.
@@ -276,7 +276,7 @@ gdb_test "record restore $precsave" \
"reload precord save file"
# Signal handlers must be re-enabled
-gdb_test "handle all stop print"
+gdb_test_lines "handle all stop print" "" ""
# Make the first signal SIGABRT because it is always supported.
set sig_supported 1
@@ -77,19 +77,23 @@ foreach inf_sel {1 2} {
gdb_test_no_output "set debug remote 1"
- gdb_test_multiple "break -q main" "set breakpoint" {
- -re "Sending packet: \\\$qXfer:auxv:read.*$gdb_prompt $" {
- # Symbol lookup may access the target to read AUXV in
- # order to determine the debug base for SVR4 linker
- # namespaces.
- xfail "$gdb_test_name"
- }
- -re "Sending packet.*$gdb_prompt $" {
- fail "$gdb_test_name"
- }
- -re "^break -q main\r\nBreakpoint .* at .*$gdb_prompt $" {
- pass "$gdb_test_name"
- }
+ set lines [gdb_get_lines "break -q main"]
+
+ set test "set breakpoint"
+ set re_xfail \
+ [string_to_regexp {Sending packet: $qXfer:auxv:read}]
+ if { [regexp $re_xfail $lines] } {
+ # Symbol lookup may access the target to read AUXV in
+ # order to determine the debug base for SVR4 linker
+ # namespaces.
+ xfail $test
+ } elseif { [regexp "Sending packet:" $lines] } {
+ fail $test
+ } else {
+ set re \
+ "Breakpoint $decimal at $hex: \[^\r\n\]+"
+ set re "^$re\r\n$"
+ gdb_assert { [regexp $re $lines] } $test
}
gdb_test_no_output "set debug remote 0"
@@ -58,8 +58,8 @@ if { [istarget *-*-linux*] } {
if { $do_xml_test } {
# This is done in a way to avoid the timeout that can occur from
# applying .* regexp to large output.
- gdb_test_sequence "info os processes" "get process list" \
- { "pid +user +command" "1 +root +\[/a-z\]*(init|systemd|bash)" }
+ gdb_test_lines "info os processes" "get process list" \
+ "^pid +user +command.*\r\n1 +root +\[/a-z\]*(init|systemd|bash)"
}
}