I've decided that the new tokenizer is here to stay and reasonably 
mature, so I thought I might just as well make it part of the official 
development branch:

https://github.com/POV-Ray/povray/releases/tag/v3.8.0-alpha.10008988

There's still quite a lot of work I intend to be doing on the main 
parser to take advantage of the new tokenizer, but for now I'll focus on 
some changes to TrueType font handling, which needs some TLC after the 
new tokenizer has shaken up charcter set handling.

Post a reply to this message

hi,

clipka <ano### [at] anonymousorg> wrote:
> I've decided that the new tokenizer is here to stay and reasonably
> mature, so I thought I might just as well make it part of the official
> development branch:
> https://github.com/POV-Ray/povray/releases/tag/v3.8.0-alpha.10008988

the code below gives me a "Segmentation fault".


regards, jr.


(from the newsgoups last year but didn't keep the reference to the author)

// Hintergrund

plane { z,0.441
 pigment { checker color rgb 0.7 color rgb 0.3 }
}

light_source {
 <-33,66,-99> color 1
 area_light <16,0,0>, <0,16,0> , 5, 5 jitter
}

camera {
 orthographic
 right <8,0,0>*0.875
 up <0,6,0>*0.875
 location <0,0,-10>
 look_at <0,0,0>
}

#declare P3 = array[8] {0, 1, 1, 1, 3,  4,  12, 9999};
#declare P4 = array[8] {0, 1, 1, 2, 5, 12,  35, 9999};
#declare P6 = array[8] {0, 1, 1, 3, 7, 22, 999, 9999};

#declare N=1;
#declare T=6;
#declare P=1;
#declare SCL = 0.2;

#declare Y=-2;
#while(Y<=+2)
 #declare X=-3;
 #while(X<=+3)

// #declare S = concat("N", str(N,2,0), ", T", str(T,2,0), ", P", str(P,2,0) );
 #declare S = concat("T", str(T,2,0), ", P", str(P,2,0) );
 #if(T<7)

// sphere { <0,0,0> , 0.45
 cylinder { <0,0,-0.45>, <0,0,+0.45>, 0.45
  pigment { pavement
    number_of_sides 4
    number_of_tiles T
    pattern P

   color_map {
    [0.0 color rgb<1.0,1.0,1.0> ]
    [1.0 color rgb<0.2,0.1,0.6> ]
   }
  rotate x*-90
  scale 0.08
  }
  finish { phong 0.5 }

  translate <X,-Y,0>
 }

 text { ttf "arialbd.ttf"
  S
  0.001 0
  pigment { color rgb<1.0, 0.6, 0.2>*1.25 }
  no_shadow
  scale SCL
  translate <X-0.4, -Y-0.45, -0.55>
 }

 text { ttf "arialbd.ttf"
  S
  0.001 0
  pigment { color rgb 0.1 }
  no_shadow
  scale SCL
  translate <X-0.4, -Y-0.45, -0.55> + 0.0081*<1,-1,1>
 }
 #end

 #declare X=X+1;
 #declare N=N+1;
 #declare P=P+1;
 #if(P>P4[T])
  #declare T=T+1;
  #declare P=1;
 #end
 #end
#declare Y=Y+1;
#end

Post a reply to this message

Am 12.01.2019 um 13:36 schrieb jr:
> the code below gives me a "Segmentation fault".

Oops, sorry about that. Until I've released a fix, you should be able to 
work around it (to some degree anyway) by using `global_setting { 
charset ascii }` (or any other value).

(Yes, I know that'll give you a deprecation warning.)

Post a reply to this message

Known issue:

Using the `text` primitive without a `charset` global setting will cause 
a hard crash. I'll release a fix as soon as I can get around to it.

Post a reply to this message

hi,

clipka <ano### [at] anonymousorg> wrote:
> Am 12.01.2019 um 13:36 schrieb jr:
> > the code below gives me a "Segmentation fault".
> Oops, sorry about that. Until I've released a fix, you should be able to
> work around it (to some degree anyway) by using `global_setting {
> charset ascii }` (or any other value).

nope.  with global_settings, and even #version for good measure, no change.

>
> (Yes, I know that'll give you a deprecation warning.)


regards, jr.

Post a reply to this message

Am 12.01.2019 um 16:59 schrieb jr:
> hi,
> 
> clipka <ano### [at] anonymousorg> wrote:
>> Am 12.01.2019 um 13:36 schrieb jr:
>>> the code below gives me a "Segmentation fault".
>> Oops, sorry about that. Until I've released a fix, you should be able to
>> work around it (to some degree anyway) by using `global_setting {
>> charset ascii }` (or any other value).
> 
> nope.  with global_settings, and even #version for good measure, no change.

It might now be segfaulting for some other entirely unrelated reason. Do 
you happen to have an end-of-file right after the `#end` (without a 
newline)?

That's another crash cause I saw on my machine after fixing the 
absence-of-charset issue; I thought the mechanism could only trigger in 
my private debug builds, but maybe I'm mistaken.

Post a reply to this message

hi,

clipka <ano### [at] anonymousorg> wrote:
> Am 12.01.2019 um 16:59 schrieb jr:
> > nope.  with global_settings, and even #version for good measure, no change.
>
> It might now be segfaulting for some other entirely unrelated reason. Do
> you happen to have an end-of-file right after the `#end` (without a
> newline)?

no.  (my editor ((g)vim) shows '[noeol]' in the status line for such files)

>
> That's another crash cause I saw on my machine after fixing the
> absence-of-charset issue; I thought the mechanism could only trigger in
> my private debug builds, but maybe I'm mistaken.


regards, jr.

Post a reply to this message

Am 13.01.2019 um 10:44 schrieb jr:
> hi,
> 
> clipka <ano### [at] anonymousorg> wrote:
>> Am 12.01.2019 um 16:59 schrieb jr:
>>> nope.  with global_settings, and even #version for good measure, no change.
>>
>> It might now be segfaulting for some other entirely unrelated reason. Do
>> you happen to have an end-of-file right after the `#end` (without a
>> newline)?
> 
> no.  (my editor ((g)vim) shows '[noeol]' in the status line for such files)

Then the bad news is that I don't know why it crashes; on my machine, 
the scene clearly bombed due to an uninitialized-variable situation in 
the `text` primitive, which shouldn't occur if you use `global_settings 
{ charset WHATEVER }` (or `#version 3.8`, actually).

The latest alpha fixes that situation. Let's see whether it also fixes 
your crashes.

If it doesn't, I'll be clinging to the straw that the crash might be 
specific to the `arialbd.ttf` font, and that I'm not seeing it because I 
might have a different version of that font on my machine; in that case 
I'd like to ask you to try with one of the distributed fonts, e.g. 
`crystal.ttf`.

Post a reply to this message

hi,

clipka <ano### [at] anonymousorg> wrote:
> Am 13.01.2019 um 10:44 schrieb jr:
> > no.  (my editor ((g)vim) shows '[noeol]' in the status line for such files)
> ...
> The latest alpha fixes that situation. Let's see whether it also fixes
> your crashes.

no crash so far, just the error (see v3.8 character set handling).

(I've already "de-commissioned" 10008988)

> If it doesn't, I'll be clinging to the straw that the crash might be
> specific to the `arialbd.ttf` font, and that I'm not seeing it because I
> might have a different version of that font on my machine; ...

not that it matters but mine is (Redmond) time-stamped "140418183001Z0", and the
file size is 750216 bytes.


regards, jr.

Post a reply to this message

Am 13.01.2019 um 14:25 schrieb jr:
> hi,
> 
> clipka <ano### [at] anonymousorg> wrote:
>> Am 13.01.2019 um 10:44 schrieb jr:
>>> no.  (my editor ((g)vim) shows '[noeol]' in the status line for such files)
>> ...
>> The latest alpha fixes that situation. Let's see whether it also fixes
>> your crashes.
> 
> no crash so far, just the error (see v3.8 character set handling).

That's good news then...

> (I've already "de-commissioned" 10008988)

... and that's an excellent decision.

Post a reply to this message