← Back to team overview

ooc-dev team mailing list archive

Re: Compiling rock


It looks like by adding this linkage flag compilation process runs flawless. I'm asking myself why with big surprise (maybe I try to get the answer from next week on). Aside from that, why not adding this flag on makefiles instead of forcing users to do it themselves?

Does compilation work on other platforms with that linker option on? If so, it would be a trivial change. If not, a check for Win32 system should be done in makefiles (it's done already, I believe, although unsure). Which are implications of having the flag?


PD: I'm unsure about erasing previous message contents on replies. What do you prefer?

-----Mensaje original----- From: Amos Wenger
Sent: Sunday, January 09, 2011 7:06 PM
To: Iván HernándezValmisa
Cc: ooc-dev@xxxxxxxxxxxxxxxxxxx
Subject: Re: [Ooc-dev] Compiling rock

Hi Ivan,

I have encountered this issue before, it's known although I have no idea
how to fix it, there's a simple work-around.

Apparently, rock gets stuck when launching ar for ExplanationGenerator.
You can work around that by just disabling lib-caching, ie. using
OOC_FLAGS=-nolibcache when building rock.

Let me know how that goes - if anyone knows why ar is getting stuck
sometime on mingw32... I have really no idea. It works fine on all Linux
and OSX flavors that I've tested.

Amos (aka nddrylliog)

On Sun, 2011-01-09 at 18:52 +0100, Iván Hernández Valmisa wrote:
Hi, everyone.

I've been absent from ooc development a few months. Now, I'm trying to
compile latest version of rock and it goes ok until gcc stalls when
compiling rock_tmp\source\rock\backend\explain

It looks like if once the source distributed with bootstrapping
package gets compiled, the bootstrapping compiler generates bad source
code for the final compiler that gcc can't digest. Maybe recursive
includes? (although gcc shouldn't have problems with this)

I'm on a Windows Seven platform with gcc 4.5.0 on msys. GIT version is
from today (January 9th), but I had the same problem in the middle of
December (I expected some waiting should correct the problem). Don't
hesitate to ask for more information, if it is needed. I will do some
tests you may suggest, but I believe it should be enough to refresh
the source code on alpaca's c_source from a recent working rock

Mailing list: https://launchpad.net/~ooc-dev
Post to     : ooc-dev@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~ooc-dev
More help   : https://help.launchpad.net/ListHelp