|
00:06
FROGGS joined
00:23
lizmat joined
02:04
ggoebel joined
05:51
FROGGS[mobile] joined
06:20
FROGGS joined
06:59
Ven joined
07:19
brrt joined
|
|||
| arnsholt | jnthn: I was going to say that implementing my patch on JVM wasn't going to be too much trouble, but then I went and looked at the JVM code and realized there's no HashAttrStore =) | 07:34 | |
| brrt | good * | 07:49 | |
|
08:27
zakharyas joined
08:53
pyrimidine joined
09:00
nebuchadnezzar joined
09:09
zakharyas joined
09:37
vendethiel joined
|
|||
| timotimo | o/ | 09:40 | |
|
10:17
btyler joined
10:19
zakharyas1 joined
10:21
oetiker joined
11:26
zakharyas joined
11:57
Ven joined
12:15
Ven joined
12:39
Ven joined
13:19
zakharyas joined
13:55
tadzik joined
14:15
JimmyZ_ joined
15:00
FROGGS joined
|
|||
| FROGGS | jnthn: ping | 15:48 | |
| jnthn: you got VisualStudioVersion in your build environnment set, right? | 16:14 | ||
| where does one read about $@ in makefiles? | 18:44 | ||
| because it does something different using gcc on windows and linux | 18:46 | ||
| timotimo | gcc doesn't do anything with makefiles, though | 18:48 | |
| i thought basically everybody uses "gmake" on windows, which is just a port of gnu make? | 18:49 | ||
| FROGGS | ahh yes | 18:52 | |
| I mean gmake of course | |||
| timotimo | :) | 18:53 | |
| FROGGS | nmake does seem to do it right though | ||
| my problem is that the target is 'dynext/libsomething.dll' and $@ is the entire target (with directory) instead of the filename | 18:54 | ||
| and that's not the case on linux or when I use nmake | |||
| timotimo | huh | 19:02 | |
| FROGGS | gcc -LC:\nqp\install/lib -lmoar -shared -Wl,--out-implib,libdynext\perl6_ops_moar.dll.a -O1 -DNDEBUG -g1 -O1 -DNDEBUG -lshell32 -lws2_32 -lmswsock -lrpcrt4 -ladvapi32 | 19:14 | |
| -lpsapi -liphlpapi -o dynext\perl6_ops_moar.dll dynext/perl6_ops.o dynext/container.o C:\nqp\install/lib/libmoar.dll.a | |||
| Can't open .lib file: libdynext\perl6_ops_moar.dll.a | |||
| and the pattern is: -Wl,--out-implib,lib$@.a | 19:15 | ||
| arnsholt | FROGGS: $@ is the name of the file currently being built | 19:29 | |
| Used in pattern rules, usually | |||
| FROGGS | including its path under gmake sadly | 19:30 | |
| arnsholt | Yeah, that's part of the file name, after all | ||
| But there's a function | |||
| $(notdir $@) should give you the filename without the directory portion | 19:31 | ||
| FROGGS | ohh | ||
| will try | |||
| arnsholt | Of course, that's GNU Make specific, which means that the Makefile will stop working with nmake | ||
| (And IIRC jnthn uses nmake =) | |||
| FROGGS | no problem there | 19:32 | |
| timotimo | "notdir"? o_O | ||
| FROGGS | it will be emited for gmake only so it is safe | 19:35 | |
| arnsholt++ # it works | 19:36 | ||
| \o/ | 19:39 | ||
| just built moar+nqp+rakudo using mingw32 that comes with strawberry perl | 19:40 | ||
|
20:07
bonsaikitten joined,
moritz joined
20:08
nwc10 joined
|
|||
| timotimo | that's new, right? | 21:26 | |
| we've been recommending visual studio instead of that, right? | |||
|
22:36
cygx joined
|
|||
| cygx | timotimo: MoarVM explicitly supports MinGW but has some minor problems with installation (the dll should to go into /bin but end up in /lib, the import lib should go into /lib but isn't copied at all) | 22:39 | |
| NQP builds fine on MinGW but tries to invoke nmake at first | 22:40 | ||
| Rakudo has issues, and I just used my own Makefile that I fixed manually: github.com/cygx/rakudo/blob/mingw3...le.mingw32 | |||
| note that on MinGW, compiling the setting fails if spesh is enabled | |||
| or at least it used to - haven't checked lately | 22:41 | ||
| o/ | 23:06 | ||
|
23:06
cygx left
|
|||