[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-readline] [PATCH] Enable visibility annotations

From: Pedro Alves
Subject: Re: [Bug-readline] [PATCH] Enable visibility annotations
Date: Fri, 22 Apr 2016 15:58:53 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101 Thunderbird/38.7.1

On 04/22/2016 03:53 PM, Chet Ramey wrote:
> On 4/22/16 10:24 AM, Pedro Alves wrote:
>>> All of the public symbols in history.h are part of the history library,
>>> which can be built both as part of readline and separately.  We were a
>>> little careless 25 years ago with the public function names, and it's too
>>> late to change those now, but all of the history library variables have a
>>> `history_' prefix.
>> Note that "history_offset" is declared in histlib.h however, which reads:
>>  /* histlib.h -- internal definitions for the history library. */
>>  ...
>>  /* Some variable definitions shared across history source files. */
>>  extern int history_offset;
> That definition is in history.h in readline-7.0.

Ah.  I started by looking at gdb's older readline copy, but
I was actually looking at the "devel" branch before sending that.
In there, we indeed have:

 history.h:extern int history_offset;

But we also still have:

 histlib.h:extern int history_offset;
 misc.c:extern int history_offset;
 examples/hist_purgecmd.c:44:extern int history_offset;
 examples/hist_erasedups.c:41:extern int history_offset;

which was what misled me.  Shouldn't all these extern
declarations except the history.h one be removed?

Pedro Alves

reply via email to

[Prev in Thread] Current Thread [Next in Thread]