git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: Patch for NO_R_TO_GCC_LINKER

From
Giovanni Funchal <gafunchal@gmail.com>
Date
Aug 13, 2008, 21:20 UTC
Message-ID
<c475e2e60808131420h5b7350abq188b3f8a53b51d98@mail.gmail.com>
In-Reply-To
<7vy730n0ac.fsf@gitster.siamese.dyndns.org>
> In this project, use of configure is strictly optional and Makefile is
> more canonical than autoconf generated configure.  Unless absolutely
> necessary, I'd prefer to have a solution that does _not_ change the set of
> make variables people need to override from the command line.

My changes should not break a lot of people's setups. Most people will find the new default better because it works straight on linux/windows. I have tested it on ubuntu gcc 4.2.3, fedora gcc 3.2.3 and sunos gcc 3.4.2 and only sunos require ./configure or command line arg. In addition this should provide better support for people running AIX, IRIX and HP-UX which, to my best knowledge, have a different way around runtime paths. Although very improbable, the changes might indeed break some setups, but keep reading...

> Changes to configure.ac so that generated script sets NO_R_TO_GCC_LINKER
> appropriately would fit the current model much better and would not break
> people's existing setups that do not use configure.

This is contradictory... how can changes ONLY to configure.ac make rpath work for people NOT using configure?

Try to face it like this: current Makefile support for rpath is broken. Defaults are bad and the NO_R_TO_GCC_LINKER is totally ugly and do not cover the case when neither "-Wl,rpath," nor "-R" are the right choice. In addition if you do not test for flags support (running configure), you cannot have better warnings when things go wrong.

Regards, -- Giovanni

On Wed, Aug 13, 2008 at 10:10 PM, Junio C Hamano <gitster@pobox.com> wrote:
Show 31 quoted lines
> "Giovanni Funchal" <gafunchal@gmail.com> writes:
>
>> diff --git a/Makefile b/Makefile
>> index 90c5a13..6e20b08 100644
>> --- a/Makefile
>> +++ b/Makefile
>> @@ -111,9 +111,8 @@ all::
>>  #
>>  # Define NO_DEFLATE_BOUND if your zlib does not have deflateBound.
>>  #
>> -# Define NO_R_TO_GCC_LINKER if your gcc does not like "-R/path/lib"
>> -# that tells runtime paths to dynamic libraries;
>> -# "-Wl,-rpath=/path/lib" is used instead.
>> +# LD_RUNPATH_SWITCH specifies how to pass the runtime dynamic library paths
>> +# to the linker. The default is "-Wl,-rpath,".
>
> Can you make your change to configure.ac to minimize changes to the
> Makefile?
>
> In this project, use of configure is strictly optional and Makefile is
> more canonical than autoconf generated configure.  Unless absolutely
> necessary, I'd prefer to have a solution that does _not_ change the set of
> make variables people need to override from the command line.
>
> Changes to configure.ac so that generated script sets NO_R_TO_GCC_LINKER
> appropriately would fit the current model much better and would not break
> people's existing setups that do not use configure.
>
>
>
>
Previous: Junio C HamanoNext: Junio C Hamano
Message 4 of 6 in “Patch for NO_R_TO_GCC_LINKER”
  1. Giovanni FunchalAug 13, 2008
  2. Matthieu MoyAug 13, 2008
  3. Junio C HamanoAug 13, 2008
  4. Giovanni FunchalAug 13, 2008
  5. Junio C HamanoAug 13, 2008
  6. Junio C HamanoAug 13, 2008

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.