Re: [Mingw-users] GCC-4.9.3 is now our current release
- Date: Sun, 20 Mar 2016 21:50:48 +0100
- From: ralph engels <ralphengels@xxxxxxxxx>
- Subject: Re: [Mingw-users] GCC-4.9.3 is now our current release
Agree it's better to investigate first why its defined in both places
and which one is correct,
if both are the same then it might be an option, but probably better
moving it to somewhere in the build chain
where its visible to all parts that use it.
"Well, the build strategy is different: you don't bootstrap a crossed
native build. I originally tried to build it, using my GCC-4.9.3 cross
compiler as the build compiler, but that didn't fly, so I bootstrapped a
native Linux GCC-5.3.0, then used that to build a GCC-5.3.0 cross
compiler; finally, used that to build the crossed native GCC-5.3.0,
(with a simple 'make all'). "
I failed to build it even if i skipped the bootstrap stage,
i did find out that it does not like being bootstrapped with a version
built against winpthreads even if i removed the winpthreads library and
headers completely the build would crash so i had to use the non posix
threads enabled gcc build from your site to bootstrap it before i got
past the crash.
Sadly i have no linux build environment handy atm. So i had to resolve
to fixing the build on my win 7 box.
Going to try and see how it reacts with the patched pthreads4w library.
Den 20-03-2016 kl. 20:31 skrev Keith Marshall:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> On 20/03/16 17:32, ralph engels wrote:
>> Aye also ran into the CONST_CAST2 warnings, but im a bit baffled
>> why it seems to build gcc on linux while not on windows.
> Well, the build strategy is different: you don't bootstrap a crossed
> native build. I originally tried to build it, using my GCC-4.9.3
> cross compiler as the build compiler, but that didn't fly, so I
> bootstrapped a native Linux GCC-5.3.0, then used that to build a
> GCC-5.3.0 cross compiler; finally, used that to build the crossed
> native GCC-5.3.0, (with a simple 'make all').
>> ill take a good look at your patches maybe something i missed :)
>> btw you can get around the CONST_CAST2 warning by undefining the
>> macro just before it is set in gcc,
> By "in gcc", I assume you mean in gcc/system.h, but then you are
> assuming that the definition there is correct, and begs the question:
> why is there an incompatible definition in mingw32/bits/gthr-default.h?
> Which is the correct definition? Are they even intended to serve the
> same purpose?
> Alternatively, maybe the definition in gthr-default.h is correct. In
> that case, the definition in system.h should be wrapped in #ifndef.
> Or maybe both are correct, in different contexts. I don't know enough
> of GCC internals to arbitrate on this; it needs to go upstream, as a
> bug report.
>> but im not sure if it breaks something else so i hold of on that
>> option for now.
>> I would be interrested seing the log, mostly for comparison with my
>> own, might give me some ideas.
> Sent privately, under separate cover; it's way too big for a list
> attachment, even when compressed. (BTW, in case you're interested in
> improving your English, there is only one "r" in "interested", and two
> "e"s in "seeing").
>> The other issues are also worth a look.
> The Wformat warnings are just worthless pedantry, but certainly the
> rest are worth a look; I'll leave it for someone who both knows about,
> and cares about Ada, to comment on the 90 "used uninitialized"
> warnings from building that component; (they are nothing new, having
> afflicted every version of GCC I've ever built, with Ada enabled).
> - --
> Public key available from keys.gnupg.net
> Key fingerprint: C19E C018 1547 DE50 E1D4 8F53 C0AD 36C6 347E 5A3F
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2.0.20 (GNU/Linux)
> -----END PGP SIGNATURE-----
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> MinGW-users mailing list
> This list observes the Etiquette found at
> We ask that you be polite and do the same. Disregard for the list etiquette may cause your account to be moderated.
> You may change your MinGW Account Options or unsubscribe at:
> Also: mailto:mingw-users-request@xxxxxxxxxxxxxxxxxxxxx?subject=unsubscribe
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
MinGW-users mailing list
This list observes the Etiquette found at
We ask that you be polite and do the same. Disregard for the list etiquette may cause your account to be moderated.
You may change your MinGW Account Options or unsubscribe at: