View previous topic :: View next topic |
Author |
Message |
catalyst1 n00b
Joined: 30 Aug 2002 Posts: 9
|
Posted: Mon Nov 18, 2002 3:30 pm Post subject: TTMKFDIR build problem on sparc 5 |
|
|
I am stuck trying to get ttmkfdir to build.
There are errors like:
/usr/include/bits/mathinline.h:216 'double fdim(double, double) throws different exceptions'
In a different thread somebody fixed this with an 'emerge flex'. This didn't help. I also emerged gcc to no avail.
On a solaris thread someone talked about an 'append patch' or something? Is there a gentoo sparc library patch? \
Thanks!
-Eric _________________ "There are more things in heaven and earth than are dreamt of in your philosophy" |
|
Back to top |
|
|
catalyst1 n00b
Joined: 30 Aug 2002 Posts: 9
|
Posted: Mon Nov 18, 2002 6:13 pm Post subject: ttmkfdir build problem |
|
|
P.S. This is a problem only with ttmkfdir 2.0. 1.0 seemed to build just fine... _________________ "There are more things in heaven and earth than are dreamt of in your philosophy" |
|
Back to top |
|
|
Weeve Retired Dev
Joined: 30 Oct 2002 Posts: 641
|
Posted: Mon Nov 18, 2002 7:05 pm Post subject: glibc is the culprit |
|
|
IIRC, glibc is the culprit here. This has been encountered before. Not sure when a fix is/was planned for Gentoo however.
Below was the patch I used to fix this. I recommend backing up your usr/include/bits/mathinline.h first, in case something happens.
To apply this patch, execute the following command
cd /usr/include/bits ; patch -p0 < /path/to/mathinline.patch
--- mathinline.h.jj 2002-11-01 12:25:55.000000000 -0500
+++ mathinline.h 2002-11-01 12:27:38.000000000 -0500
@@ -213,14 +213,14 @@
# ifndef __NO_MATH_INLINES
-__MATH_INLINE double fdim (double __x, double __y);
+__MATH_INLINE double fdim (double __x, double __y) __THROW;
__MATH_INLINE double
fdim (double __x, double __y)
{
return __x < __y ? 0 : __x - __y;
}
-__MATH_INLINE float fdimf (float __x, float __y);
+__MATH_INLINE float fdimf (float __x, float __y) __THROW;
__MATH_INLINE float
fdimf (float __x, float __y)
{ |
|
Back to top |
|
|
catalyst1 n00b
Joined: 30 Aug 2002 Posts: 9
|
Posted: Mon Nov 18, 2002 10:52 pm Post subject: Re: glibc is the culprit |
|
|
Thanks!
I changed the 4.2.1 Xfree ebuild to rely on V 1.0 of ttmkfdir instead. It's building right now... I guess no TT fonts would be the worst thing that would happen???? We'll see.
So, mathlinline.patch is a modified copy of mathinline.h? With the following parts changed? I'm not entirely sure what I'm looking at here.
Thanks again!
-Cat1
# ifndef __NO_MATH_INLINES
-__MATH_INLINE double fdim (double __x, double __y);
+__MATH_INLINE double fdim (double __x, double __y) __THROW;
__MATH_INLINE double
fdim (double __x, double __y)
{
return __x < __y ? 0 : __x - __y;
}
-__MATH_INLINE float fdimf (float __x, float __y);
+__MATH_INLINE float fdimf (float __x, float __y) __THROW;
__MATH_INLINE float
fdimf (float __x, float __y)
{ _________________ "There are more things in heaven and earth than are dreamt of in your philosophy" |
|
Back to top |
|
|
catalyst1 n00b
Joined: 30 Aug 2002 Posts: 9
|
Posted: Mon Nov 18, 2002 10:55 pm Post subject: I think I got it.... |
|
|
I just read about patch. Seems pretty straightforward.
I'll see what happens with this bastardized 4.2.1 build and let you know.
Cheers,
-Cat1 _________________ "There are more things in heaven and earth than are dreamt of in your philosophy" |
|
Back to top |
|
|
savatoge n00b
Joined: 19 Nov 2002 Posts: 3
|
Posted: Tue Nov 19, 2002 3:16 pm Post subject: Same problem here... |
|
|
I have installed Gentoo 1.4_rc1 on an Ultra10...emerging x11base/xfree is still giving me the same error message as describe in the first message in this thread, even after I applied the patch to mathinline.sh.
Am I missing something here? |
|
Back to top |
|
|
savatoge n00b
Joined: 19 Nov 2002 Posts: 3
|
Posted: Tue Nov 19, 2002 4:02 pm Post subject: Re: glibc is the culprit |
|
|
I found the problem...the patch is missing the added __THROW in two places...should be:
# ifndef __NO_MATH_INLINES
-__MATH_INLINE double fdim (double __x, double __y);
+__MATH_INLINE double fdim (double __x, double __y) __THROW;
__MATH_INLINE double
-fdim (double __x, double __y)
+ fdim (double __x, double __y) __THROW
{
return __x < __y ? 0 : __x - __y;
}
-__MATH_INLINE float fdimf (float __x, float __y);
+__MATH_INLINE float fdimf (float __x, float __y) __THROW;
__MATH_INLINE float
- fdimf (float __x, float __y)
+fdim (double __x, double __y) __THROW
{
Thanks for the posts...got me pointed in the right direction. xfree is building as I type. |
|
Back to top |
|
|
Weeve Retired Dev
Joined: 30 Oct 2002 Posts: 641
|
Posted: Tue Nov 26, 2002 6:45 pm Post subject: |
|
|
Thanks for the clarification |
|
Back to top |
|
|
|