Re: Quantifying the time overhead of Cygwin make

2014-06-16 Thread Michael Stahl
On 10/06/14 13:06, Michael Stahl wrote: On 10/06/14 12:47, Bjoern Michaelsen wrote: Hi, On Tue, Jun 10, 2014 at 01:22:16PM +0300, Tor Lillqvist wrote: You mean cmd.exe + coreutils should be enough? Probably bash or some other POSIX shell too, surely? Yeah. Note that e.g. busybox already

Re: Quantifying the time overhead of Cygwin make

2014-06-12 Thread Jonathan Aquilina
I am not trying to hijack this thread, but I am part of another project where we build for windows on linux boxes. Have you guys considered looking into building for windows on Linux? On Wed, Jun 11, 2014 at 1:37 PM, Norbert Thiebaud nthieb...@gmail.com wrote: On Tue, Jun 10, 2014 at 4:43 AM,

Re: Quantifying the time overhead of Cygwin make

2014-06-12 Thread Stephan Bergmann
On 06/12/2014 08:30 AM, Jonathan Aquilina wrote: I am not trying to hijack this thread, but I am part of another project where we build for windows on linux boxes. Have you guys considered looking into building for windows on Linux? We have support for mingw cross-compilation in the build

Re: Quantifying the time overhead of Cygwin make

2014-06-12 Thread Jonathan Aquilina
What would be interesting to see is a bench mark in terms of compilation times and evaluate if its really worth changing the ABI and breaking compatability. On Thu, Jun 12, 2014 at 8:50 AM, Stephan Bergmann sberg...@redhat.com wrote: On 06/12/2014 08:30 AM, Jonathan Aquilina wrote: I am not

Re: Quantifying the time overhead of Cygwin make

2014-06-12 Thread Tor Lillqvist
What would be interesting to see is a bench mark in terms of compilation times and evaluate if its really worth changing the ABI and breaking compatability. Of course it isn't. Imagine telling a customer yes, we know the binary extensions that are business critical to you and which you

Re: Quantifying the time overhead of Cygwin make

2014-06-11 Thread Norbert Thiebaud
On Tue, Jun 10, 2014 at 4:43 AM, Tor Lillqvist t...@iki.fi wrote: Are you sure GnuWin32 actually is something even worth considering? Cygwin at least has a semi-decent package management even if it is a bit slower. If I understand correctly the exercise is about replacing gmake itself from a

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Bjoern Michaelsen
Hi, On Mon, Jun 09, 2014 at 11:47:41PM +0200, Michael Stahl wrote: IV. Conclusion Using a native Win32 GNU make provides faster from-scratch rebuilds than any previous attempt to improve build performance, and significantly reduces the incremental re-build overhead of make on Windows by a

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Tor Lillqvist
Are you sure GnuWin32 actually is something even worth considering? Cygwin at least has a semi-decent package management even if it is a bit slower. --tml ___ LibreOffice mailing list LibreOffice@lists.freedesktop.org

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Bjoern Michaelsen
Hi, On Tue, Jun 10, 2014 at 12:43:05PM +0300, Tor Lillqvist wrote: Are you sure GnuWin32 actually is something even worth considering? Cygwin at least has a semi-decent package management even if it is a bit slower. IMHO, the build system shouldnt go much beyond coreutils with its expectation

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Tor Lillqvist
IMHO, the build system shouldnt go much beyond coreutils with its expectation on the baseline, You mean cmd.exe + coreutils should be enough? Probably bash or some other POSIX shell too, surely?. Is there a non-Cygwin such that actually would support all the shell constructs we use? Is its

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Bjoern Michaelsen
Hi, On Tue, Jun 10, 2014 at 01:22:16PM +0300, Tor Lillqvist wrote: You mean cmd.exe + coreutils should be enough? Ideally yes, plus some POSIX shell. Probably bash or some other POSIX shell too, surely? Yeah. Note that e.g. busybox already includes the ash shell. Is there a non-Cygwin such

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Michael Stahl
On 10/06/14 12:47, Bjoern Michaelsen wrote: Hi, On Tue, Jun 10, 2014 at 01:22:16PM +0300, Tor Lillqvist wrote: You mean cmd.exe + coreutils should be enough? Ideally yes, plus some POSIX shell. and i want a pony. there's a long list of Cygwin packages that are currently needed, and

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Tor Lillqvist
it's possible that Msys has a lot of the things we need, Technically MSYS is a fork of Cygwin, though. Faster? Assume so, yes, why else would they have bothered. More immune to the Imagine a future Windows version breaking cygwin fundamentally effect? No idea. I did use MSYS quite a lot some

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Michael Stahl
On 10/06/14 11:04, Bjoern Michaelsen wrote: In the long run we might add our local patches to the native make again, if its worth it (maybe?), but that can be done incrementally. Also incrementally, we maybe they would provide some additional speedup, although i never liked the built-in cp

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Bjoern Michaelsen
Hi, On Tue, Jun 10, 2014 at 01:06:11PM +0200, Michael Stahl wrote: setup-x86.exe ... autoconf automake pkg-config Yeah, those are a pain to port (ironically -- as their original purpose was to provide portability -- to now obsolete systems). I was only half joking when I considered CMake to

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Stephan Bergmann
On 06/10/2014 01:36 PM, Bjoern Michaelsen wrote: OTOH, as said above, some of those might just end up in external/ then (or already are) ... ...except for those needed by external itself (into which category probably falls wget) Stephan ___

Re: Quantifying the time overhead of Cygwin make

2014-06-10 Thread Bjoern Michaelsen
Hi, On Tue, Jun 10, 2014 at 02:21:04PM +0200, Stephan Bergmann wrote: On 06/10/2014 01:36 PM, Bjoern Michaelsen wrote: OTOH, as said above, some of those might just end up in external/ then (or already are) ... ...except for those needed by external itself (into which category probably