#195 svn r5723 does not build

[Stefan Husmann]

The newest commit seems to cause an interruption of the compilation of the sources.
This happens using texinfo 5.2 and 6+ from svn.

[Charles Staats]

Three notes:
1) I have reverted the change until the issue can be tracked down. Please verify that r5724 builds for you.
2) I was unable to duplicate your error.
3) As closely as I can tell, the error that interrupted your compilation occured on line 2 of the asymptote.texi file, which was not changed by r5723.

Thus, I am somewhat mystified.

One thought: It looks like you are only recompiling files that have changed. What happens if you recompile r5722 from scratch, and compare it to r5723 recompiled from scratch? (Possible explanation: one of my changes was to add a new asy file that needed to be compiled. Difficulties in compiling other asy files might not have been noticed earlier, since the asy files themselves did not change.)

[Stefan Husmann]

Am 01.08.2015 um 19:09 schrieb Charles Staats:

Three notes:
1) I have reverted the change until the issue can be tracked down. Please verify that r5724 builds for you.
2) I was unable to duplicate your error.
3) As closely as I can tell, the error that interrupted your compilation occured on line 2 of the asymptote.texi file, which was not changed by r5723.

Thus, I am somewhat mystified.

One thought: It looks like you are only recompiling files that have changed. What happens if you recompile r5722 from scratch, and compare it to r5723 recompiled from scratch? (Possible explanation: one of my changes was to add a new asy file that needed to be compiled. Difficulties in compiling other asy files might not have been noticed earlier, since the asy files themselves did not change.)

---

[bugs:#195] svn r5723 does not build

Hello,
I compiled r5724 successfully from scratch.

I am quite sure I did so also for r5724, but will recheck.

Edit: r5723 fails again, but with a different error.

Thanks for your help.

Best regards
Stefan Husmann

[Charles Staats]

I thought I responded to this several hours ago, but apparently it didn't take. Sorry about the delay.

Anyway, here was the thrust: the error you are showing is strange, as if genustwo.asy were not present in your copy of r5723. Please verify that it's there (it should be in the doc folder) and see what happens if you order a compilation manually. From the doc folder, the command to compile it is

../asy -dir ../base -config "" -render=0 -f pdf -noprc genustwo.asy

Also a general note: It appears as if the top part of the output from your make commands is getting cut off. The text files you are giving would be even more useful if you were to save the entire output to a file, say by appending &>makeoutput.log when you issue the command.

[Stefan Husmann]

Hello,

Am 02.08.2015 um 04:11 schrieb Charles Staats:

../asy -dir ../base -config "" -render=0 -f pdf -noprc genustwo.asy

The file eists and the given command states:

[haawda@frege doc]$ ../asy -dir ../base -config "" -render=0 -f pdf -noprc genustwo.asy
../base/three.asy: 2905.13: runtime: division by 0 in transform of a triple
[haawda@frege doc]$ echo $?
0

There is no new file generated.

Best Regards

Stefan Husmann

[Charles Staats]

Now I'm less mystified: r5723 did not actually cause a bug; rather, it brought an old issue to the fore by inadvertantly adding an additional test in the form of genustwo.asy.

I suspect your rendering engine is having trouble producing rasterized output; see https://sourceforge.net/p/asymptote/bugs/162/. To test this theory, try running within the doc folder

../asy -dir ../base -config "" -render=4 -f pdf -noprc GaussianSurface.asy

and see if you get the same error as for genustwo.asy.

[Stefan Husmann]

Am 03.08.2015 um 02:07 schrieb Charles Staats:

../asy -dir ../base -config "" -render=4 -f pdf -noprc GaussianSurface.asy
and see if you get the same error as for genustwo.asy.

Hello,

I get
../base/three.asy: 2905.13: runtime: division by 0 in transform of a triple

for both GaussianSurface.asy and genustwo.asy, and this in both revisions r5723 and r5724.

I have freeglut 3.0.0 and glu 9.0.0, both from the official Arch Linux repos. Are there better-working flavours known?

Thanks again and best regards
Stefan

[Charles Staats]

You're asking the wrong person. See the three posts beginning with
https://sourceforge.net/p/asymptote/discussion/409349/thread/9e80ba77/#05d2.
If that doesn't help, you may need to try to get John Bowman involved.

In the meantime, I'm working on a hack that may be able to replace the
image with a text box saying "Rasterized image failed to render" rather
than interrupting the build.

On Mon, Aug 3, 2015 at 8:58 AM, Stefan Husmann haawda@users.sf.net wrote:

Am 03.08.2015 um 02:07 schrieb Charles Staats:

../asy -dir ../base -config "" -render=4 -f pdf -noprc GaussianSurface.asy
and see if you get the same error as for genustwo.asy.

Hello,

I get
../base/three.asy: 2905.13: runtime: division by 0 in transform of a triple

for both GaussianSurface.asy and genustwo.asy, and this in both revisions
r5723 and r5724.

I have freeglut 3.0.0 and glu 9.0.0, both from the official Arch Linux
repos. Are there better-working flavours known?

Thanks again and best regards
Stefan

---

• [bugs:#195] http://sourceforge.net/p/asymptote/bugs/195 svn r5723 does
  not build*

Status: open
Group: v1.0 (example)
Created: Sat Aug 01, 2015 03:32 PM UTC by Stefan Husmann
Last Updated: Mon Aug 03, 2015 12:07 AM UTC
Owner: Charles Staats
Attachments:

• bug.txt
  http://sourceforge.net/p/asymptote/bugs/195/attachment/bug.txt (16.1
  kB; text/plain)

The newest commit seems to cause an interruption of the compilation of the
sources.
This happens using texinfo 5.2 and 6+ from svn.

---

Sent from sourceforge.net because you indicated interest in
https://sourceforge.net/p/asymptote/bugs/195/

To unsubscribe from further messages, please visit
https://sourceforge.net/auth/subscriptions/

[Stefan Husmann]

It seems this is a bug in the propiatary nvidia driver toolchain. With nouveau drivers it vanished.