Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Simon Marchi <simark@simark.ca>
To: Sergio Durigan Junior <sergiodj@redhat.com>,
	GDB Patches <gdb-patches@sourceware.org>
Cc: Simon Marchi <simon.marchi@ericsson.com>
Subject: Re: [PATCH 1/2] Create new common/pathstuff.[ch]
Date: Sun, 11 Feb 2018 22:14:00 -0000	[thread overview]
Message-ID: <084a12bf-6b54-9399-e4fe-887952a57bc1@simark.ca> (raw)
In-Reply-To: <20180210014241.19278-2-sergiodj@redhat.com>

On 2018-02-09 08:42 PM, Sergio Durigan Junior wrote:
> This commit moves the path manipulation routines found on utils.c to a
> new common/pathstuff.c, and updates the Makefile.in's accordingly.
> The routines moved are "gdb_realpath", "gdb_realpath_keepfile" and
> "gdb_abspath".
> 
> This will be needed because gdbserver will have to call "gdb_abspath"
> on my next patch, which implements a way to expand the path of the
> inferior provided by the user in order to allow specifying just the
> binary name when starting gdbserver, like:
> 
>   $ gdbserver :1234 a.out
> 
> With the recent addition of the startup-with-shell feature on
> gdbserver, this scenario doesn't work anymore if the user doesn't have
> the current directory listed in the PATH variable.
> 
> I had to do a minor adjustment on "gdb_abspath" because we don't have
> access to "tilde_expand" on gdbserver, so now the function is using
> "gdb_tilde_expand" instead.  Otherwise, the code is the same.
> 
> Regression tested on the BuildBot, without regressions.

Hi Sergio,

Thanks for looking into this!

This commit does not build:

/home/simark/src/binutils-gdb/gdb/common/pathstuff.c: In function ‘gdb::unique_xmalloc_ptr<char> gdb_abspath(const char*)’:
/home/simark/src/binutils-gdb/gdb/common/pathstuff.c:140:14: error: ‘current_directory’ was not declared in this scope
     (concat (current_directory,
              ^~~~~~~~~~~~~~~~~
/home/simark/src/binutils-gdb/gdb/common/pathstuff.c:140:14: note: suggested alternative: ‘read_direction’
     (concat (current_directory,
              ^~~~~~~~~~~~~~~~~
              read_direction


I guess you need to move the declaration to common-defs.h in this commit instead of
the next one.  I also got this whitespace error from git am:

.git/rebase-apply/patch:131: trailing whitespace.
  /* Extract the basename of filename, and return immediately

I think it's in code you copied, but if you can remove the extra space while at it
it would be nice.

> +/* See common/pathstuff.h.  */
> +
> +gdb::unique_xmalloc_ptr<char>
> +gdb_abspath (const char *path)
> +{
> +  gdb_assert (path != NULL && path[0] != '\0');
> +
> +  if (path[0] == '~')
> +    {
> +      std::string new_path = gdb_tilde_expand (path);
> +
> +      return gdb::unique_xmalloc_ptr<char> (xstrdup (new_path.c_str ()));

We should try to avoid unnecessary copies, when possible.  Here, we could either make
another version of gdb_tilde_expand (it would have to be another name) that returns
a gdb::unique_xmalloc_ptr<char> or make gdb_abspath return an std::string.  I think the
former would be better for now because some callers require a gdb::unique_xmalloc_ptr<char>,
and would have to do a copy themselves.  So using a gdb::unique_xmalloc_ptr across the
whole chain would give the least amount of copies.

> diff --git a/gdb/utils.h b/gdb/utils.h
> index b234762929..4f25be0968 100644
> --- a/gdb/utils.h
> +++ b/gdb/utils.h
> @@ -23,6 +23,7 @@
>  
>  #include "exceptions.h"
>  #include "common/scoped_restore.h"
> +#include "common/pathstuff.h"

I don't think utils.h should be including common/pathstuff.h, because it is not using it.
I understand why you did this (this ensures that every current user of these functions
will automatically see the new declaration), but in the long term I think it's better if
the users include "pathstuff.h".  The files that use these functions are:

$ grep -e gdb_realpath_keepfile -e gdb_realpath -e gdb_abspath *.c */*.c -l | sort
auto-load.c
common/pathstuff.c
compile/compile.c
dwarf2read.c
exec.c
guile/scm-safe-call.c
linux-thread-db.c
main.c
nto-tdep.c
objfiles.c
source.c
symtab.c
utils.c

They are all files included in a --enable-targets=all build, so it should be easy to
find where #include "common/pathstuff.h" is missing.

Thanks,

Simon


  reply	other threads:[~2018-02-11 22:14 UTC|newest]

Thread overview: 51+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-02-10  1:42 [PATCH 0/2] Make gdbserver work with filename-only binaries Sergio Durigan Junior
2018-02-10  1:42 ` [PATCH 2/2] " Sergio Durigan Junior
2018-02-12  4:18   ` Simon Marchi
2018-02-12 19:16     ` Sergio Durigan Junior
2018-02-21  8:05       ` Joel Brobecker
2018-02-12 19:57   ` [PATCH 0/2] " Sergio Durigan Junior
2018-02-12 19:57     ` [PATCH v2 1/2] Create new common/pathstuff.[ch] Sergio Durigan Junior
2018-02-12 19:57     ` [PATCH v2 2/2] Make gdbserver work with filename-only binaries Sergio Durigan Junior
2018-02-13  4:35       ` Simon Marchi
2018-02-22 18:37         ` Sergio Durigan Junior
2018-02-21 12:29       ` Pedro Alves
2018-02-27  0:20         ` Sergio Durigan Junior
2018-02-28  3:32           ` Sergio Durigan Junior
2018-02-28  3:27   ` [PATCH v3 0/2] " Sergio Durigan Junior
2018-02-28  3:27     ` [PATCH v3 2/2] " Sergio Durigan Junior
2018-02-28  3:58       ` Sergio Durigan Junior
2018-02-28  5:33         ` Simon Marchi
2018-02-28  7:09           ` Metzger, Markus T
2018-02-28 16:30             ` Sergio Durigan Junior
2018-02-28  5:46       ` Simon Marchi
2018-02-28 16:29         ` Sergio Durigan Junior
2018-02-28 16:40           ` Sergio Durigan Junior
2018-02-28  3:27     ` [PATCH v3 1/2] Create new common/pathstuff.[ch] Sergio Durigan Junior
2018-02-28  5:02       ` Simon Marchi
2018-02-28 16:46         ` Sergio Durigan Junior
2018-02-28 16:39       ` Sergio Durigan Junior
2018-03-01  2:23     ` [PATCH v3 0/2] Make gdbserver work with filename-only binaries Sergio Durigan Junior
2018-03-01  2:55       ` Joel Brobecker
2018-03-01 13:08         ` Christophe Lyon
2018-03-01 13:18           ` Simon Marchi
2018-03-01 19:50           ` Sergio Durigan Junior
2018-03-01 20:20           ` [PATCH] Conditionally include "<windows.h>" on common/pathstuff.c (and unbreak build on mingw*) Sergio Durigan Junior
2018-03-01 20:47             ` Simon Marchi
2018-03-02 11:46               ` Christophe Lyon
2018-03-02 12:35                 ` Sergio Durigan Junior
2018-03-02 11:11             ` Yao Qi
2018-03-02 12:29               ` Sergio Durigan Junior
2018-03-02 12:37                 ` Sergio Durigan Junior
2018-03-05 12:07                   ` Yao Qi
2018-03-02 13:32             ` Eli Zaretskii
2018-03-02 15:15               ` Simon Marchi
2018-03-02 18:20                 ` Sergio Durigan Junior
2018-03-03  7:36                   ` Eli Zaretskii
2018-03-01 17:37         ` [PATCH v3 0/2] Make gdbserver work with filename-only binaries Sergio Durigan Junior
2018-03-02  3:20           ` Joel Brobecker
2018-02-28 16:47   ` [obvious/pushed] Change order of error message printed when gdbserver can't find CWD Sergio Durigan Junior
2018-02-10  1:42 ` [PATCH 1/2] Create new common/pathstuff.[ch] Sergio Durigan Junior
2018-02-11 22:14   ` Simon Marchi [this message]
2018-02-12 19:01     ` Sergio Durigan Junior
2018-02-21  7:56   ` Joel Brobecker
2018-02-22 18:43     ` Sergio Durigan Junior

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=084a12bf-6b54-9399-e4fe-887952a57bc1@simark.ca \
    --to=simark@simark.ca \
    --cc=gdb-patches@sourceware.org \
    --cc=sergiodj@redhat.com \
    --cc=simon.marchi@ericsson.com \
    /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