Hi Stuart, On Sat, Apr 07, 2018 at 08:35:10PM +0100, Stuart Henderson wrote: > My dev machines are updated past clang 6 so I can't test atm but it reads > ok.
We might as well fix the build first. The below diff should do the trick. I *think* I'm right to not bump. OK for that part? > I don't think it will do much to bulk build speeds in itself, coccinelle > and tex-by-topic aren't parts of big dependency chains, and don't take > that long to build, but seems a good testbed for your script which could > well be useful for things in future that have a bigger impact on the > build. Agreed. Thanks Index: patches/patch-texk_web2c_luatexdir_image_pdftoepdf_w =================================================================== RCS file: patches/patch-texk_web2c_luatexdir_image_pdftoepdf_w diff -N patches/patch-texk_web2c_luatexdir_image_pdftoepdf_w --- /dev/null 1 Jan 1970 00:00:00 -0000 +++ patches/patch-texk_web2c_luatexdir_image_pdftoepdf_w 8 Apr 2018 10:38:34 -0000 @@ -0,0 +1,17 @@ +$OpenBSD$ + +Fix C++11 build problem in generated C++ code. +http://www.tug.org/pipermail/tex-live/2017-June/040365.html + +Index: texk/web2c/luatexdir/image/pdftoepdf.w +--- texk/web2c/luatexdir/image/pdftoepdf.w.orig ++++ texk/web2c/luatexdir/image/pdftoepdf.w +@@ -71,7 +71,7 @@ static char *get_file_checksum(const char *a, file_err + ck = (char *) malloc(PDF_CHECKSUM_SIZE); + if (ck == NULL) + formatted_error("pdf inclusion","out of memory while processing '%s'", a); +- snprintf(ck, PDF_CHECKSUM_SIZE, "%" PRIu64 "_%" PRIu64, (uint64_t) size,(uint64_t) mtime); ++ snprintf(ck, PDF_CHECKSUM_SIZE, "%"@= @>PRIu64@= @>"_%"@= @>PRIu64, (uint64_t) size,(uint64_t) mtime); + } else { + switch (fe) { + case FE_FAIL: -- Best Regards Edd Barrett http://www.theunixzoo.co.uk
