From: Bruno Larsen <blarsen@redhat.com>
To: Pedro Alves <pedro@palves.net>, gdb-patches@sourceware.org
Subject: Re: [PATCH v4 2/4] gdb/cli: add '.' as an argument for 'list' command
Date: Mon, 17 Jul 2023 10:21:10 +0200 [thread overview]
Message-ID: <430fb2d9-ae07-6551-3426-adb5334c127d@redhat.com> (raw)
In-Reply-To: <835af40f-edae-ba48-f121-a1cdd1f1147e@palves.net>
On 14/07/2023 19:50, Pedro Alves wrote:
> Hi,
>
> Sorry for coming late to the party here, but while trying to catch up on the
> list I spotted a few things that I think should be fixed. See below.
I forgot to send the email, but I've already pushed this patch... I'll
fix up the things you pointed in a moment and send a new patch, I just
have a few questions.
>
> BTW, I think the feature is cool!
>
> On 2023-07-13 11:24, Bruno Larsen via Gdb-patches wrote:
>> Currently, after the user has used the list command once, there is no
>> self-contained way to ask GDB to print the location where the inferior is
>> stopped. The current best options require either using a separate
>> command to scope out where the inferior is stopped, or using "list *$pc"
>> requiring knowledge of GDB standard registers. This commit adds a way
>> to do that using '.' as a new argument for the 'list' command. If the
>> inferior isn't running, the command prints around the main function.
>>
>> Because this necessitated having the inferior running and the test was
>> (seemingly unnecessarily) using printf in a non-essential way and it
>> would make the resulting log harder to read for no benefit, it was
>> replaced by a differen t statement.
>> ---
>> gdb/NEWS | 4 ++++
>> gdb/cli/cli-cmds.c | 31 ++++++++++++++++++++++++--
>> gdb/doc/gdb.texinfo | 5 +++++
>> gdb/testsuite/gdb.base/list.exp | 39 +++++++++++++++++++++++++++++++++
>> gdb/testsuite/gdb.base/list1.c | 2 +-
>> 5 files changed, 78 insertions(+), 3 deletions(-)
>>
>> diff --git a/gdb/NEWS b/gdb/NEWS
>> index b924834d3d7..eef440a5242 100644
>> --- a/gdb/NEWS
>> +++ b/gdb/NEWS
>> @@ -84,6 +84,10 @@
>> is 64k. To print longer strings you should increase
>> 'max-value-size'.
>>
>> +* The 'list' command now accepts '.' as an argument, which tells GDB to
>> + print the location where the inferior is stopped. If the inferior hasn't
>> + started yet, the command will print around the main function.
> It would be more accurate to say that it's where the current thread is stopped.
>
> Say you run to breakpoint hit in thread 1, and then switch to thread 2, and then do
> "list .". That will show you the current frame of thread 2, while "where the
> inferior is stopped" could very well be interpreted as "thread 1".
The wording does need changing, I agree. I think using the wording that
already exists in the documentation is the best way to go: "prints
around the last solitary line printed as part of displaying a stack
frame". The wording I originally used (and the improvement you
suggested) could make it sound like I would always print the lowermost
frame, when that is not the point I wanted. What I wanted was to re-do
what the first usage of "list" does without needing to know the lines.
My big question: Can I change the NEWS entry, or is that some taboo
that's best avoided?
>
>> +
>> * New commands
>>
>> maintenance print record-instruction [ N ]
>> diff --git a/gdb/cli/cli-cmds.c b/gdb/cli/cli-cmds.c
>> index 00977bc2ee3..1c459afdc97 100644
>> --- a/gdb/cli/cli-cmds.c
>> +++ b/gdb/cli/cli-cmds.c
>> @@ -1234,14 +1234,14 @@ list_command (const char *arg, int from_tty)
>> const char *p;
>>
>> /* Pull in the current default source line if necessary. */
>> - if (arg == NULL || ((arg[0] == '+' || arg[0] == '-') && arg[1] == '\0'))
>> + if (arg == NULL || ((arg[0] == '+' || arg[0] == '-' || arg[0] == '.') && arg[1] == '\0'))
>> {
>> set_default_source_symtab_and_line ();
>> symtab_and_line cursal = get_current_source_symtab_and_line ();
>>
>> /* If this is the first "list" since we've set the current
>> source line, center the listing around that line. */
>> - if (get_first_line_listed () == 0)
>> + if (get_first_line_listed () == 0 && (arg == nullptr || arg[0] != '.'))
>> {
>> list_around_line (arg, cursal);
>> }
>> @@ -1263,6 +1263,32 @@ list_command (const char *arg, int from_tty)
>> print_source_lines (cursal.symtab, range, 0);
>> }
>>
>> + /* "l ." lists the default location again. */
> Spelling out "list ." would be better for grepping, IMHO.
I followed the convention of the code around it. All used only 'l'
instead of "list", but I can change that easily enough.
>
>> + else if (arg[0] == '.')
>> + {
>> + try
>> + {
>> + /* Find the current line by getting the PC of the currently
>> + selected frame, and finding the line associated to it. */
>> + frame_info_ptr frame = get_selected_frame (nullptr);
> So this is actually using the selected frame, not the current frame. So even
> the "where the thread is stopped" description would be incorrect. If you really
> wanted to use frame #0 (where the thread stopped), then this should use get_current_frame.
Yeah... that's the confusion I mentioned earlier... I want the selected
frame, since its what "list" does when you first call it after entering
a frame.
>
>
>> + CORE_ADDR curr_pc = get_frame_pc (frame);
>> + cursal = find_pc_line (curr_pc, 0);
>> + }
>> + catch (const gdb_exception &e)
>> + {
>> + /* If there was an exception above, it means the inferior
>> + is not running, so reset the current source location to
>> + the default. */
> Aww. Please don't use a try/catch for this... You can just check target_has_execution.
Oh, I thought target_has_execution would query if the target can execute
the inferior, rather than looking if the inferior is being executed.
Will change
>
> Also, if an exception would be good for this (which it isn't), please don't catch
> gdb_exception -- it catches QUIT/Ctrl-C as well, which is most often wrong. You would
> want gdb_exception_error normally.
Also news to me! I probably won't use it, but its good to know for when
I need it :)
>> + clear_current_source_symtab_and_line ();
>> + set_default_source_symtab_and_line ();
>> + cursal = get_current_source_symtab_and_line ();
>> + }
>> + list_around_line (arg, cursal);
>> + /* Advance argument so just pressing "enter" after using "list ."
>> + will print the following lines instead of the same lines again. */
>> + arg++;
>> + }
>> +
>> return;
>> }
>>
>> @@ -2770,6 +2796,7 @@ and send its output to SHELL_COMMAND."));
>> = add_com ("list", class_files, list_command, _("\
>> List specified function or line.\n\
>> With no argument, lists ten more lines after or around previous listing.\n\
>> +\"list .\" lists ten lines arond where the inferior is stopped.\n\
> arond -> around
>
>> \"list -\" lists the ten lines before a previous ten-line listing.\n\
>> One argument specifies a line, and ten lines are listed around that line.\n\
>> Two arguments with comma between specify starting and ending lines to list.\n\
>> diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
>> index b10c06ae91f..7619efe3de9 100644
>> --- a/gdb/doc/gdb.texinfo
>> +++ b/gdb/doc/gdb.texinfo
>> @@ -9148,6 +9148,11 @@ Stack}), this prints lines centered around that line.
>>
>> @item list -
>> Print lines just before the lines last printed.
>> +
>> +@item list .
>> +Print the lines surrounding the location that is where the inferior
>> +is stopped. If the inferior is not running, print around the main
>> +function instead.
> This should be clarified as well.
>
>> @end table
>>
>> @cindex @code{list}, how many lines to display
>> diff --git a/gdb/testsuite/gdb.base/list.exp b/gdb/testsuite/gdb.base/list.exp
>> index 18ecd13ac0f..ed178a1dd95 100644
>> --- a/gdb/testsuite/gdb.base/list.exp
>> +++ b/gdb/testsuite/gdb.base/list.exp
>> @@ -400,6 +400,42 @@ proc test_list_invalid_args {} {
>> "second use of \"list +INVALID\""
>> }
>>
>> +proc test_list_current_location {} {
>> + global binfile
>> + # If the first "list" command that GDB runs is "list ." GDB may be
>> + # unable to recognize that the inferior isn't running, so we should
>> + # reload the inferior to test that condition.
> I don't understand this comment. Why would it be unable to do so?
doh, this is a comment form previous iterations. Back when I thought
"list ." should error, we were entering the "never before listed" if
clause, instead of finding "list ." which was a problem. Now that the
behavior is the same, there isn't a reason to restart the inferior.
>
>> + clean_restart
>> + gdb_file_cmd ${binfile}
>> +
>> + # Ensure that we are printing 10 lines
>> + if {![set_listsize 10]} {
>> + return
>> + }
>> +
>> + # First guarantee that GDB prints around the main function correctly
>> + gdb_test "list ." \
>> + "1.*\r\n2\[ \t\]+\r\n3\[ \t\]+int main \[)(\]+.*5\[ \t\]+int x;.*" \
>> + "list . with inferior not running"
>> +
>> + if {![runto_main]} {
>> + warning "couldn't start inferior"
>> + return
>> + }
>> +
>> + # Walk forward some lines
> Missing period.
>
>> + gdb_test "until 15" ".*15.*foo.*"
>> +
>
>> + # Test that the correct location is printed and that
>> + # using just "list" will print the following lines.
>> + gdb_test "list ." ".*" "list current line after starting"
>> + gdb_test "list" ".*" "confirm we are printing the following lines"
>> +
>> + # Test that list . will reset to current location
>> + gdb_test "list ." ".*" "list around current line again"
>> + gdb_test " " ".*" "testing repeated invocations with GDB's auto-repeat"
> I don't understand -- these 4 tests match ".*", so how how they ensuring what
> they claim they test? Looks like WIP?
you are correct, it is WIP... taking time off while developing this
patch was clearly a bad choice hahaha
--
Cheers,
Bruno
>
> Pedro Alves
>
>> +}
>> +
>> clean_restart
>> gdb_file_cmd ${binfile}
>>
>> @@ -519,4 +555,7 @@ test_list "list -" 10 2 "7-8" "5-6"
>> # the current line.
>> test_list "list -" 10 1 "7" "6"
>>
>> +# Test printing the location where the inferior is stopped.
>> +test_list_current_location
>> +
>> remote_exec build "rm -f list0.h"
>> diff --git a/gdb/testsuite/gdb.base/list1.c b/gdb/testsuite/gdb.base/list1.c
>> index d694495c3fb..9297f958f46 100644
>> --- a/gdb/testsuite/gdb.base/list1.c
>> +++ b/gdb/testsuite/gdb.base/list1.c
>> @@ -7,7 +7,7 @@ void bar (int x)
>> -
>> - */
>> {
>> - printf ("%d\n", x);
>> + x++;
>>
>> long_line ();
>> }
next prev parent reply other threads:[~2023-07-17 8:21 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-07-13 10:24 [PATCH v4 0/4] Small changes to "list" command Bruno Larsen
2023-07-13 10:24 ` [PATCH v4 1/4] gdb/cli: Factor out code to list lines around a given line Bruno Larsen
2023-07-13 16:53 ` Tom Tromey
2023-07-13 10:24 ` [PATCH v4 2/4] gdb/cli: add '.' as an argument for 'list' command Bruno Larsen
2023-07-13 11:05 ` Eli Zaretskii
2023-07-13 16:53 ` Tom Tromey
2023-07-14 17:50 ` Pedro Alves
2023-07-17 8:21 ` Bruno Larsen [this message]
2023-07-17 8:44 ` Andrew Burgess
2023-07-17 14:14 ` Pedro Alves
2023-07-18 11:21 ` [PATCH] gdb/cli: fixes to newly added "list ." command Bruno Larsen
2023-07-18 12:54 ` Eli Zaretskii
2023-07-18 13:40 ` Bruno Larsen
2023-07-18 16:17 ` Eli Zaretskii
2023-07-18 13:43 ` Pedro Alves
2023-07-18 14:55 ` Bruno Larsen
2023-07-21 10:26 ` [PATCH v2] " Bruno Larsen
2023-07-21 11:05 ` Eli Zaretskii
2023-08-04 8:37 ` [PING][PATCH " Bruno Larsen
2023-08-23 10:03 ` [PINGv2][PATCH " Guinevere Larsen
2023-08-23 15:00 ` [PATCH " Andrew Burgess
2023-08-28 15:50 ` [PATCH v3] " Guinevere Larsen
2023-09-14 13:00 ` [PING][PATCH " Guinevere Larsen
2023-09-18 13:16 ` [PATCH " Andrew Burgess
2023-09-19 9:06 ` [PATCH v4] " Guinevere Larsen
2023-09-19 11:27 ` Eli Zaretskii
2023-09-19 12:07 ` Guinevere Larsen
2023-07-13 10:24 ` [PATCH v4 3/4] gdb/cli: Improve UX when using list with no args Bruno Larsen
2023-07-13 11:06 ` Eli Zaretskii
2023-07-13 17:41 ` Keith Seitz
2023-07-13 10:24 ` [PATCH v4 4/4] gdb/doc: document '+' argument for 'list' command Bruno Larsen
2023-07-13 17:35 ` Keith Seitz
2023-07-13 21:30 ` Matt Rice
2023-07-14 8:53 ` Bruno Larsen
2023-07-14 16:30 ` Tom Tromey
2023-07-14 21:30 ` Matt Rice
2023-07-13 17:31 ` [PATCH v4 0/4] Small changes to "list" command Tom Tromey
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=430fb2d9-ae07-6551-3426-adb5334c127d@redhat.com \
--to=blarsen@redhat.com \
--cc=gdb-patches@sourceware.org \
--cc=pedro@palves.net \
/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).