This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] Speed up "gdb -tui" output
- From: Doug Evans <dje at google dot com>
- To: Eli Zaretskii <eliz at gnu dot org>
- Cc: Pedro Alves <palves at redhat dot com>, gdb-patches <gdb-patches at sourceware dot org>
- Date: Wed, 7 Jan 2015 10:09:03 -0800
- Subject: Re: [PATCH] Speed up "gdb -tui" output
- Authentication-results: sourceware.org; auth=none
- References: <83zj9v7urq dot fsf at gnu dot org> <CADPb22Q7oD3K-dYkngEPDBbV++mLCKifTEmvJczQ=0h2FX0yXA at mail dot gmail dot com> <54AD4E22 dot 1010106 at redhat dot com> <83mw5u79t7 dot fsf at gnu dot org>
On Wed, Jan 7, 2015 at 9:57 AM, Eli Zaretskii <eliz@gnu.org> wrote:
>> Date: Wed, 07 Jan 2015 15:17:54 +0000
>> From: Pedro Alves <palves@redhat.com>
>> CC: gdb-patches <gdb-patches@sourceware.org>
>>
>> > Reading symbols from foo...done.
>> > mumble ...
>> > (gdb)
>>
>> Since stdout is line buffered by default (on Unix), if this is
>> working when the TUI is disabled, then it must be because there's
>> explicit gdb_flush(gdb_stdout) after "Reading symbols from foo..."
>> is printed, right?
>
> Yes.
>
>> Isn't the issue then that the TUI's implementation of
>> gdb_flush (tui/tui-file.c) should be doing whatever it
>> needs to flush the output?
>
> Maybe. I already made that change in my sandbox. Alas, that's just
> the tip of the iceberg. We call single-character output functions,
> like putchar and fputc all over the place, implicitly assuming that in
> certain cases (e.g., when the stream is stderr), things are unbuffered
> and the output appears immediately. On top of that, functions that
> output strings go through single-character output versions, so there's
> no easy way of saying on, say, utils.c level whether the character
> should or shouldn't appear immediately. It's a terrible mess.
>
>> Should it be calling wrefresh if the file is gdb_stdout?
>
> Only if that stream is a real file, not a string or whatever else we
> support in ui-file.
tui_file_flush only gets called for tui files, so no worries there.
Though I see tui also has its own ui file strings (tui_sfileopen).
That can go I think.
>
>> If we do that, and change tui_puts like:
>>
>> - /* We could defer the following. */
>> - wrefresh (w);
>> - fflush (stdout);
>> + if (c == '\n')
>> + gdb_flush (gdb_stdout);
>>
>> would it work?
>
> No, it's not enough. There's also gdb_stderr and gdb_stdlog, at
> least. (tui_puts doesn't get the stream as its argument, so it
> doesn't really know which stream it is serving.) I'm trying to make
> heads or tails out of this, but I'm not there yet.
Sure, but it does suggest tui_puts is the wrong place to do
any kind of flushing/refreshing.
Fortunately, tui_puts is internal to tui.