Follow-up Comment #5, bug #67997 (group groff):

At 2026-02-04T18:08:48-0500, Dave wrote:
> Follow-up Comment #4, bug #67997 (group groff):
>
> [comment #3 comment #3:]
>> Look for dot files in that directory.
>
> I have, and I also use rmdir, which fails if the directory has any
> contents.

Ah.

>> If my half-informed speculations are wrong and the stuff in
>> "_Inline" _really is_ deleted every time the Perl interpreter exits,
>
> I don't know about every time in general use, but certainly every time
> for my (admittedly limited) gropdf use.

Just a wild guess, but maybe compilation of the object files is lazy,
but creation of the "inline" directory is not.  So if _gropdf_ doesn't
need to do the thing that it builds C code to optimize, it never gets
compiled in the first place.

I think it has to do with font subsetting on huge fonts.

Checking...

Ah, well, first thing I see is that "gropdf.pl" _does not_ create the
directory "lazily", but non-contingently on whether XS will be used.


 248 mkdir $ENV{HOME}.'/_Inline' if !-e $ENV{HOME}.'/_Inline' and
!exists($ENV{PERL_INLINE_DIRECTORY}) and exists($ENV{HOME});


...but...


 250 $rc = eval
 251 {
 252     require Inline;
 253     Inline->import (C => Config => DIRECTORY => $ENV{HOME}.'/_Inline') if
!exists($ENV{PERL_INLINE_DIRECTORY}) and exists($ENV{HOME});
 254     Inline->import (C =><<'EOC');
 255
 256     static const uint32_t MAGIC1 = 52845;
 257     static const uint32_t MAGIC2 = 22719;
 258
 259     typedef unsigned char byte;
 260
 261     char* decrypt_exec_C(char *s, int len)
 262     {
 263         static uint16_t er=55665;
 264         byte clr=0;
 265         int i;
 266         er=55665;
 267
 268         for (i=0; i < len; i++)
 269         {
 270             byte cypher = s[i];
 271             clr = (byte)(cypher ^ (er >> 8));
 272             er = (uint16_t)((cypher + er) * MAGIC1 + MAGIC2);
 273             s[i] = clr;
 274         }
 275
 276         return(s);
 277     }
 278
 279 EOC
 280 };


...that doesn't look very contingent, either.

Beats me.



    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?67997>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/

Attachment: signature.asc
Description: PGP signature

Reply via email to