[tlbuild] initial test builds for TL'15 (no commits)

[Mojca Miklavec]

On Tue, Mar 17, 2015 at 12:46 AM, Akira Kakuto wrote:
> Hi Karl,
>
>> Anyway, I changed the test to use "sed 1,2d", which should work
>> everywhere (until future standards "improve" sed).
>
>
> I suspect that attached
> xetex-bug73.test.diff-1
> or
> xetex-bug73.test.diff-2
> seem necessary.
> Sorry if I am wrong.
>
> Thanks,
> Akira

Thanks a lot for fixing this. Now I ran into a new problem:

================================================
   ttfdump (TeX Live) 0.5.5: ./test-suite.log
================================================

# TOTAL: 1
# PASS:  0
# SKIP:  0
# XFAIL: 0
# FAIL:  1
# XPASS: 0
# ERROR: 0

.. contents:: :depth: 2

FAIL: tests/ttfdump
===================

*** ttfdump -t glyf -g 1 -o ttftest.glyf FonetikaDaniaIwonaeRegular.ttf
ttfdump.glyf tests OK

*** ttfdump -t GPOS -o ttftest.gpos FonetikaDaniaIwonaeRegular.ttf
ttfdump.gpos tests OK

*** ttfdump -t GSUB -o ttftest.gsub FonetikaDaniaIwonaeRegular.ttf
ttfdump.gsub tests OK

*** ttfdump -t head FonetikaDaniaIwonaeRegular.ttf >ttftest.head
ttfdump.head tests OK

*** ttfdump -t OS/2 texmfhome.otf >ttftest.os_2
ttfdump.os_2 tests OK

*** ttfdump -t OS/2 bad.otf >ttftest.bad
ttfdump.bad tests OK

*** ttfdump -t gasp -c 0 test.ttc >ttftest.gasp
Segmentation Fault - core dumped
*** ttfdump -t maxp -c 1 test.ttc >ttftest.maxp
Segmentation Fault - core dumped
failed tests: ttfdump.gasp ttfdump.maxp

----------------------------------

I manually copied test.ttc to source/Work/texk/ttfdump/ because I
couldn't get the kpathsea settings right and used "./test.ttc".

> gdb ./ttfdump
(gdb) r -t gasp -c 0 ./test.ttc >ttftest.gasp
Starting program: [...]/Build/source/Work/texk/ttfdump/ttfdump -t gasp
-c 0 ./test.ttc >ttftest.gasp
[Thread debugging using libthread_db enabled]
[New Thread 1 (LWP 1)]
[Switching to Thread 1 (LWP 1)]
warning: kpathsea: configuration file texmf.cnf not found in these
directories: /../kpathsea.

Program received signal SIGSEGV, Segmentation fault.
0xff242d10 in strlen () from /lib/libc.so.1
(gdb) backtrace
#0  0xff242d10 in strlen () from /lib/libc.so.1
#1  0xff2af3f0 in _ndoprnt () from /lib/libc.so.1
#2  0xff2b1420 in fprintf () from /lib/libc.so.1
#3  0x0002f40c in print_prologue (out=0x48b10 <_iob+16>) at
../../../texk/ttfdump/src/ttfdump.c:327
#4  main (argc=<optimized out>, argv=<optimized out>) at
../../../texk/ttfdump/src/ttfdump.c:178
(gdb) b 327
Breakpoint 1 at 0x2f3f4: file ../../../texk/ttfdump/src/ttfdump.c, line 327.
(gdb) r -t gasp -c 0 ./test.ttc >ttftest.gasp
The program being debugged has been started already.
Start it from the beginning? (y or n) y
Starting program: [...]/Build/source/Work/texk/ttfdump/ttfdump -t gasp
-c 0 ./test.ttc >ttftest.gasp
[Thread debugging using libthread_db enabled]
[New Thread 1 (LWP 1)]
[Switching to Thread 1 (LWP 1)]

warning: kpathsea: configuration file texmf.cnf not found in these
directories: /../kpathsea.

Breakpoint 1, print_prologue (out=0x48b10 <_iob+16>) at
../../../texk/ttfdump/src/ttfdump.c:327
327       fprintf(out, "Dumping File:%s\n\n\n", font->ttfname);
(gdb) p font
$1 = (TTFontPtr) 0x57f90
(gdb) p font->ttfname
$2 = 0x0
(gdb)




Independent:

I have just noticed that I still have the following patch from last
year in the svn checkout:

--- source/texk/xdvik/configure (revision 36544)
+++ source/texk/xdvik/configure (working copy)
@@ -15996,6 +15996,7 @@
 # Checks for iconv

 # Check whether --with-iconv was given.
+with_iconv=no
 if test "${with_iconv+set}" = set; then :
   withval=$with_iconv;
 else

I added that as a quick workaround to avoid linking against libiconv
which isn't present on the system by default. What would be the best
way to handle this dependency? Running ./Build --without-iconv (or
something similar)?

Mojca