automake-patches
[Top][All Lists]
Advanced

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

Re: Fixes for Windows


From: Mathieu Lirzin
Subject: Re: Fixes for Windows
Date: Fri, 30 Mar 2018 15:06:08 +0200
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/25.3 (gnu/linux)

Hello Olly,

Sorry for the delay.

Olly Betts <address@hidden> writes:

> On 2018-01-24, Olly Betts <address@hidden> wrote:
>> I found I needed the two attached patches to build Xapian
>> (https://xapian.org) on Windows.
>>
>> The second patch is only needed when using libtool, but should
>> be harmless when not.
>
> It's now been two months since I sent these, and there's been no
> response at all and it doesn't look like they've been applied, despite
> two new automake releases being made since.  
>
> Is this list no longer the appropriate place to send patches?
> If not, then README needs updating.

This is the appropriate place for sending patches.

The main reason that your patch didn't get reviewed yet is that I am
incompetent for issues related to Cygwin/MinGW/MSYS, so I was waiting
for someone to jump in and then forgot about it.  :-)

> From 91fb05a87c7a9eb784a84b14b74313740189c402 Mon Sep 17 00:00:00 2001
> From: Olly Betts <address@hidden>
> Date: Wed, 24 Jan 2018 13:38:03 +1300
> Subject: [PATCH 1/2] Use cygpath for filename conversion on MSYS
>
> * lib/ar-lib(func_file_conv): Set file_conv=cygwin for MSYS.
> * lib/compile(func_file_conv): Likewise.
> ---

This one seems harmless.

> From 010511fe83a765238e0cb050621beee5a00fce69 Mon Sep 17 00:00:00 2001
> From: Olly Betts <address@hidden>
> Date: Wed, 24 Jan 2018 13:43:42 +1300
> Subject: [PATCH 2/2] lib/compile: Handle libtool .lo files
>
> * lib/compile(func_cl_wrapper): Add *.lo to list of files to handle
> as object files.
> ---
>  lib/compile | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib/compile b/lib/compile
> index 7851f7c4..e2931cf9 100755
> --- a/lib/compile
> +++ b/lib/compile
> @@ -143,7 +143,7 @@ func_cl_wrapper ()
>         # configure might choose to run compile as 'compile cc -o foo foo.c'.
>         eat=1
>         case $2 in
> -         *.o | *.[oO][bB][jJ])
> +         *.o | *.lo | *.[oO][bB][jJ])
>             func_file_conv "$2"
>             set x "$@" -Fo"$file"
>             shift

Unless I am mistaken, this issue is more general and applies every time
‘compile’ is used for ‘.lo’ files.  I think a test revealing the issue
can be added in the test suite. You can either modify an existing one
from the ‘t/compile*.sh’ files if the issue is closely related to an
already tested one, or if that is not the case you can add a new test
script.

Can you add a rationale in the commit message before the ChangeLog part
too?

Thanks.

-- 
Mathieu Lirzin
GPG: F2A3 8D7E EB2B 6640 5761  070D 0ADE E100 9460 4D37



reply via email to

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