apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "William A. Rowe, Jr." <wr...@rowe-clan.net>
Subject Re: [vote] release apr 1.3.4, apr-util 1.3.5
Date Tue, 02 Jun 2009 04:00:11 GMT
Guenter Knauf wrote:
> Hi,
> William A. Rowe, Jr. schrieb:
>> carlo.bramix wrote:
>>> Unfortunately, you are right: these sources do not work under Mingw and Msys
(sigh...)
>>>
>>> >From      : "Guenter Knauf" fuankg@apache.org
>>>> MingW32 / MSYS configure build is still broken ...
>>>>
>>>> G√ľnter.
>> Posts like this get you no where; post the errors and your own diagnostics
>> of them before even asking this list for help with such things.  Shame.
> oh, I think he did well since he reminded you for an outstanding patch
> before release:
> https://issues.apache.org/bugzilla/show_bug.cgi?id=46175

All that is required when reporting a problem is ...

"Does not compile" followed by one of...

  * see bug X
  * the compile errors are ...

Saying "it's broke" is not helpful, even when the bug is already reported.
I'm sorry, the post lacks context.

> I cant currently check if the patch solves all issues since I have no
> Python on my Win32 machine yet where I have MingW32/MSYS, but I can say
> that I had already a partly similar patch in pipeline before Carlo
> pointed  to the BZ. Also I doubt that Mladen broke this since main prob
> is configure magic to detect and include tlhelp32.h, and AFAIK Mladen
> does only MSVC (though maybe that changed).

Uhm - he added tlhelp32.h in support of threadproc/win32/proc.c (and had
not actually needed to add it to apr.hw, so its moved).

> Unfortunately cross-mingw32 on Linux is broken either for another
> reason: configure wants to check somewhere for /dev/zero:
> ...
> configure:23216: result: no
> configure:23230: checking for MAP_ANON in sys/mman.h
> conftest.c:35:22: error: sys/mman.h: No such file or directory
> configure:23259: result: no
> configure:23270: checking for /dev/zero
> configure:23276: error: cannot check for file existence when cross compiling
> 
> if interested the complete config.log is here:
> http://svwe10.itex.at/~hacki/config_log.txt.gz

Well, _RUN is invalid, so this is as-expected.  Perhaps it should be fixed,
but I was testing on windows so didn't see this flaw, and we can spend the
next couple weeks working through these.

It sure helped to give up on the notion of ever running autoconf on mingw :)
The dtk pkgs and alternatives haven't been updated in 6 yrs.  Guessing this
is a lost cause.



Mime
View raw message