On Wednesday, 21 January 2015 at 01:24:35 UTC, ketmar via
Digitalmars-d wrote:
On Wed, 21 Jan 2015 01:07:03 +0000
"Adam D. Ruppe via Digitalmars-d" <digitalmars-d@puremagic.com>
wrote:
On Wednesday, 21 January 2015 at 01:04:11 UTC, ketmar via
Digitalmars-d wrote:
> two days making dmd.exe work with Wine to write the simple
> utility which i can write in two hours using MinGW
> cross-compiler.
I've been using dmd with wine for a long time... I find it
sometimes seems to deadlock, but when it does, I just ctrl+c
and run it again. Mildly annoying but not a showstopper.
for me it segfaults constantly. being unable to debug it i
first made
it compilable with MinGW, and then found that root/async.c
is... well,
the root of all problems. i don't even want to know how it
works and
what it tries to accelerate; there is non-threaded version
there, so i
just made my build use it and WOW! no more crashes. the most
painful
task was to build the working dmd.exe with mingw.
to be honest, i first tried to build it with visual studio,
thinking
that it's something that dmc does wrong, and that alone took me
half of
a day (vs build scripts aren't working, so i have to struggle
with that
too in additional to vs itself). then i took a false start and
wasted
another day. and then i stopped before trying to turn dmd into
cross-compiler. ;-)
I had some nice experiment control/visualization software in D
that was making good impressions (not that physicists really care
about programming languages, but it ran fast, had a pleasant
syntax and I could safely make major reconfigurations in really
short timespans, which made people notice) but ran into a nasty
bug in some National Instruments drivers for Linux (which NI
doesn't appear to be too interested in fixing) and then wasn't
able to build DMD git-head on 64-bit Windows 7 after a couple of
weeks of trying. It was like a slow-motion train wreck.
So now I use LabView and hate everything about it.