unofficial mirror of notmuch@notmuchmail.org
 help / color / mirror / code / Atom feed
From: Jani Nikula <jani@nikula.org>
To: Vladimir Marek <Vladimir.Marek@Oracle.COM>,
	Notmuch Mail <notmuch@notmuchmail.org>
Subject: Re: [PATCH 4/4] Explicitly type void* pointers
Date: Tue, 10 Apr 2012 00:31:34 +0300	[thread overview]
Message-ID: <87mx6ka8y1.fsf@nikula.org> (raw)
In-Reply-To: <20120409181543.GC10554@pub.czech.sun.com>

Vladimir Marek <Vladimir.Marek@Oracle.COM> writes:

> Hi,
>
>> Hi, does notmuch not compile without this? IIRC talloc_steal is a macro
>> that's supposed to provide type safety (at least with GCC), and I'd be
>> hesitant about adding the casts. Please look in your talloc.h.
>
> It does not compile. It might be that I'm using Sun/Oracle CC instead of
> gcc. The error looks like this:
>
> "lib/database.cc", line 1368: Error: Cannot assign void* to const char*.

In general, that's not a difference in the C++ compilers. You can't
assign 'void *' to 'T *' in C++.

> When looking into talloc documentation, the definition seems to be:
>
> void* talloc_steal ( const void * new_ctx, const void * ptr )
>
> http://talloc.samba.org/talloc/doc/html/group__talloc.html#gaccc66139273e727183fb5bdda11ef82c
>
>
> When looking into talloc.h, it says:
>
> /* try to make talloc_set_destructor() and talloc_steal() type safe,
>    if we have a recent gcc */

It just so happens that the trick for type safety fixes the problem for
recent GCC by having an explicit cast.

> So, maybe the way to satisfy everyone would be:
>
> notmuch_message_t *tmp_message = message;
> talloc_steal(notmuch, tmp_message);
> return(tmp_message);
>
> Or alternatively,
>
> #if (__GNUC__ >= 3)
>        return talloc_steal (notmuch, message);
> #else
>        return (notmuch_message_t*) talloc_steal (notmuch, message);
> #fi
>
>
> Of course I'm happy either way :)

I'm throwing in a third alternative below. Does it work for you? I think
it's both prettier and uglier than the above at the same time! ;)

A middle ground would be to change the callers to use
"notmuch_talloc_steal", and just #define notmuch_talloc_steal
talloc_steal if __GNUC__ >= 3.

One could argue upstream talloc should have this, but OTOH it's a C
library.

BR,
Jani.


diff --git a/lib/notmuch-private.h b/lib/notmuch-private.h
index ea836f7..83b46e8 100644
--- a/lib/notmuch-private.h
+++ b/lib/notmuch-private.h
@@ -499,4 +499,22 @@ _notmuch_filenames_create (const void *ctx,
 
 NOTMUCH_END_DECLS
 
+#ifdef __cplusplus
+/* Implicit typecast from 'void *' to 'T *' is okay in C, but not in
+ * C++. In talloc_steal, an explicit cast is provided for type safety
+ * in some GCC versions. Otherwise, a cast is required. Provide a
+ * template function for this to maintain type safety, and redefine
+ * talloc_steal to use it.
+ */
+#if !(__GNUC__ >= 3)
+template <class T>
+T *notmuch_talloc_steal(const void *new_ctx, const T *ptr)
+{
+    return static_cast<T*>(talloc_steal(new_ctx, ptr));
+}
+#undef talloc_steal
+#define talloc_steal notmuch_talloc_steal
+#endif
+#endif
+
 #endif

  reply	other threads:[~2012-04-09 21:31 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-04-09 10:17 notmuch on Solaris Vladimir.Marek
2012-04-09 10:17 ` [PATCH 1/4] Make configure use /bin/bash instead of /bin/sh Vladimir.Marek
2012-04-09 11:10   ` Jani Nikula
2012-04-09 12:19     ` Vladimir Marek
2012-04-10 17:26       ` Tomi Ollila
2012-04-11  8:43         ` Vladimir Marek
2012-04-11 18:47           ` Tomi Ollila
2012-04-30 11:58   ` David Bremner
2012-04-30 12:09     ` Tomi Ollila
2012-04-09 10:17 ` [PATCH 2/4] dirent->d_type not available on Soalris Vladimir.Marek
2012-04-09 11:17   ` Jani Nikula
2012-04-09 15:12     ` Vladimir Marek
2012-04-09 15:46       ` Adam Wolfe Gordon
2012-04-09 16:32         ` Vladimir.Marek
2012-04-11 18:57           ` Tomi Ollila
2012-04-11 19:36           ` Austin Clements
2012-04-09 10:17 ` [PATCH 3/4] Private strsep implementation Vladimir.Marek
2012-10-15  5:05   ` Ethan Glasser-Camp
2012-04-09 10:17 ` [PATCH 4/4] Explicitly type void* pointers Vladimir.Marek
2012-04-09 11:04   ` Jani Nikula
2012-04-09 18:15     ` Vladimir Marek
2012-04-09 21:31       ` Jani Nikula [this message]
2012-04-10  6:53         ` Vladimir Marek
2012-04-11 21:11         ` Austin Clements
2012-04-12  8:02           ` Jani Nikula
2012-04-12 17:57             ` Austin Clements
2012-04-15 21:05   ` Jani Nikula

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

  List information: https://notmuchmail.org/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87mx6ka8y1.fsf@nikula.org \
    --to=jani@nikula.org \
    --cc=Vladimir.Marek@Oracle.COM \
    --cc=notmuch@notmuchmail.org \
    /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.
Code repositories for project(s) associated with this public inbox

	https://yhetil.org/notmuch.git/

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).