Re: DLL export with mingw-w64: currently a no-op - Mailing list pgsql-hackers

From Andrew Dunstan
Subject Re: DLL export with mingw-w64: currently a no-op
Date
Msg-id 4DC16CBD.1000506@dunslane.net
Whole thread Raw
In response to Re: DLL export with mingw-w64: currently a no-op  ("Johann 'Myrkraverk' Oskarsson" <johann@2ndquadrant.com>)
Responses Re: DLL export with mingw-w64: currently a no-op  ("Johann 'Myrkraverk' Oskarsson" <johann@2ndquadrant.com>)
List pgsql-hackers

On 05/04/2011 01:25 AM, Johann 'Myrkraverk' Oskarsson wrote:
> On Tue, 03 May 2011 12:40:28 -0000, Andrew Dunstan 
> <andrew@dunslane.net> wrote:
>
>> Our Makefiles use dlltool and dllwrap to create DLLs. If you used our 
>> recommended build method pgxs would do lots of the work for you.
>>
>> I'm not sure why you asked the mingw-w64 team about building a 
>> Postgres extension - they are unlikely to know anything much about 
>> our build system.
>
> However, they do know about the mingw build tools.  In particular, from:
>   http://oldwiki.mingw.org/index.php/dllwrap
>
> "dllwrap is a tool to build DLLs. It seems to be deprecated in favour of
> gcc -shared option, but some projects still use it. SQLite, for example."
>
> Armed with this information, it may be prudent review the DLL build
> process in PGXS.
>
> For the record: I originally asked the mingw-w64 team for help to link
> against the 64bit JVM.DLL for which there was no corresponding .def file.
>
>
> Google cache:
>   
>
http://webcache.googleusercontent.com/search?q=cache:tZkweT_cNQQJ:oldwiki.mingw.org/index.php/dllwrap+mingw+dllwrap&cd=1&hl=is&ct=clnk&gl=is&client=opera&source=www.google.is


OK, but if we want to get rid of using dllwrap we'd need a complete 
patch for it. Just changing the definition of the macro without changing 
the rest isn't going to cut it, I think. But it might well be worth doing.

cheers

andrew


pgsql-hackers by date:

Previous
From: Mitsuru IWASAKI
Date:
Subject: patch for new feature: Buffer Cache Hibernation
Next
From: Tom Lane
Date:
Subject: Re: Extreme bloating of intarray GiST indexes