William Bader | 10 Feb 01:37 2016

patch to epstopdf


I attached a small patch to epstopdf that adds a "-gray" option to produce a B&W PDF.

I get EPS files from various sources that I need to print without triggering a color separation.



--- epstopdf-2015	2016-01-18 03:23:20.428744975 +0100
+++ epstopdf-2015-wb	2016-01-18 03:39:10.653914869 +0100
 <at>  <at>  -203,6 +203,7  <at>  <at> 
 $::opt_embed = 1;
 $::opt_exact = 0;
 $::opt_filter = 0;
 $::opt_gs = 1;
 $::opt_gscmd = "";
  <at> ::opt_gsopt = ();
 <at>  <at>  -363,6 +364,7  <at>  <at> 
   --(no)debug        output debugging info  (default: $bool[$::opt_debug])
   --(no)exact        scan ExactBoundingBox  (default: $bool[$::opt_exact])
   --(no)filter       read standard input    (default: $bool[$::opt_filter])
+  --(no)gray         force gray             (default: $bool[$::opt_gray])
   --(no)gs           run ghostscript        (default: $bool[$::opt_gs])
   --(no)hires        scan HiResBoundingBox  (default: $bool[$::opt_hires])

 <at>  <at>  -431,6 +433,7  <at>  <at> 
+  "gray!",
   "gscmd=s",              # \ref{val_gscmd}
 <at>  <at>  -585,6 +588,8  <at>  <at> 

 push  <at> GS, '-dUseFlateCompression=false' unless $::opt_compress;

+push  <at> GS, qw(-sColorConversionStrategy=Gray -dProcessColorModel=/DeviceGray) if $::opt_gray;
 if ($::opt_res and
     not $::opt_res =~ /^(\d+(x\d+)?)$/) {
   warnerr "Invalid resolution: $opt_res";
Andreas Scherer | 1 Feb 11:51 2016

[CWEB] Cleanup release 3.64ae

Dear all,

a few days ago, DEK released 'CWEB 3.64ae' and directed support questions to 
your address. Following is a list of issues I'd like to see addressed:

* Do not pack backup files, ctangled sources, and executables in the archive;
  this concerns the 'examples/' subdirectory.

* The expanded 'ctangle.c' somehow doesn't match its origin 'ctangle.w';
  it uses '<stdlib.h>' instead of predeclaring 'string' functions and it is
  unclear, where this (manual?) change comes from.

* 'wordtest.w' should define 'PHICLONE' as 'unsigned' by appending a 'u'
  to the numeric value to avoid a compiler warning.

A multi-decade project to improve and extend CWEB can be found at 
https://github.com/ascherer/cwebbin – it contains far too many details to list 
here, but at least on Linux (both RPM and DEB) it makes use of the unmodified 
source drop 'cweb-3.64ae.tar.gz' and applies its stuff on top.

Andreas Scherer

Bertram Scharpf | 25 Jan 18:50 2016

Bug report for The TeXbook

Dear TeX maintainers,

I suppose I found a minor bug in The TeXbook.

For the item '\newhelp' the index directs to pages
'346, +347'. In my printed version (I bought it around 2002)
and, as far as I see, in the current version the '\newhelp'
macro only appears on page 347. The definition even comes
before the description.

Thanks in advance for processing this report and eventually
for fixing the bug.



Bertram Scharpf
Bopserstraße 15, D-70180 Stuttgart
Ruf 0711/2599337, Fax 0711/2184509                   [13*79*2531,prim]

Igor Spectorsky | 5 Jan 23:30 2016

Bug in dvipdfmx with including eps


Many thanks for creating dvipdfm, I've been using this utility since 
about 1996. However, when installing Windows 10, I've faced thew problem 
translating the dvi with eps graphics. The utility reports like that: 
filtering file via command -->mgs.exe.....
I think I've recognized the place with bug: %o% in the command string 
for calling mgs.exe expanded into something like 
c:/users/myname/AppDataLocal/Temp//dvipdfm-x..... - there is double"/". 
  However, I couldn't find proper config option to eliminate one of "/". 
I suspect that this string is formed in .exe, and I think dpxfile.c is 
responding for that. However,  I'm afraid to change this rather 
complicated source code because it may lead to other error.
So, if it is possible, please help me to fix the problem.

Thanks in advance

PS My version of MiKTeX is from setup-2.9.5721-x64.exe

Dr. Brian Biswell | 19 Oct 02:07 2015

cannot call bibtex with absolute path name

I am trying to use bibtex on a document. I have TeXLive 2015 installed on Mac OS X El Capitan.  I am running it from the Terminal. When I call bibtex with either an absolute path name or relative path name with ‘/./‘ in it, I get an error saying it cannot write to the .blg file.

The following examples will work on my machine:

bibtex texfile
bibtex path/to/texfile

the following will not work

bibtex /abs/path/to/texfile
bibtex path/to/./texfile
bibtex ../texfile

I get the same error in TeXLive 2014, but TeXLive 2011 seems to work properly (the absolute and relative paths will work properly).

Felix Pahl | 17 Oct 18:05 2015

misleading ls-R advice

Hi -- I followed the instructions at https://tug.org/texinfohtml/kpathsea.html to use the command

ls -LAR ./ >ls-R

to generate a ls-R file. However, this didn't work, and debugging with -kpathsea-debug=32 displayed messages of the form

kdebug:db:match(/usr/local/texlive/2015basic/texmf-local//tex/latex/lastpage/lastpage.sty,/usr/local/texlive/2015basic/texmf-local/tex/latex//) = 0

where the double slash prevented the match -- the correpsonding entry in the ls-R file was


The problem was resolved when I used

ls -LAR . >ls-R

instead (without the extra slash) -- the entry in the ls-R file then becomes


(again without the extra slash), and the match succeeds:

kdebug:db:match(/usr/local/texlive/2015basic/texmf-local/tex/latex/lastpage/lastpage.sty,/usr/local/texlive/2015basic/texmf-local/tex/latex//) = 1

Norbert Preining | 18 Sep 02:51 2015

portability of format dumps

Dear all

I would like to know what is ther current status of format dump file's
portability between architectures.
(status: texlive 2015 sources)

I remember faintly that at some point it was possible to use 
format dumps within the same endian-class, but it was not possible
to switch to a different-endian system.

Is this still the case, are there any other restrictions concerning
format dump reuse across platforms?

Background: Debian uses multi-arch systems and would like to have as many
packages marked up as arch-independent so that one can use whatever arch
that is installed



PREINING, Norbert                               http://www.preining.info
JAIST, Japan                                 TeX Live & Debian Developer
GPG: 0x860CDC13   fp: F7D8 A928 26E3 16A1 9FA0  ACF0 6CAC A448 860C DC13

Bruno Voisin | 19 Aug 14:32 2015

List of file type NAMEs in the TL doc

Hi Karl (or whoever monitors this address),

Regarding the list of names for --show-path, §7.2.2 of the texlive doc in TEXMFDIST/doc/texlive, that is


refers to --format which refers in turn to "kpsewhich --help", yielding

-format=NAME           use file type NAME (see list below).
-show-path=NAME        output search path for file type NAME (list below).

There is no list below, unfortunately, just

Email bug reports to tex-k@...
Kpathsea home page: http://tug.org/kpathsea/

Actually the list may be found in §§5.5.1 and 6.1 of the kpathsea doc in TEXMFDIST/doc/kpathsea, that is 


but it seems it's been overlooked at some point when updating the texlive and kpathsea docs.

Or am I missing something?

Bruno Voisin

Pavel Sanda | 14 Aug 06:22 2015

epstopdf bug?


attached is eps file which loads fine when using several postscript viewers
(e.g. gv, evince) but the content inside frame disappears when converting via

The problem can be consistently reproduced by using matlab imagesc function
for graphs with scaled x axis. Not sure whether it's matlab producing
broken eps or epstopdf problem, but given that several viewers have no 
problem to display the file I guess the later is the case.

Can anyone reproduce or give a hint how to fix this problem?

Attachment (2015-03-14.eps): application/postscript, 155 KiB
jfbu | 1 May 19:50 2015

kpsewhich --var-value


the kpathsea manual says 


•  ‘--var-value=variable’ outputs the value of variable, expanding ‘$’ (see Variable
expansion and ‘~’ (see Tilde expansion) constructs, but not performing other expansions.


$ env TEXMFHOME="/tmp" kpsewhich --var-value '$TEXMFHOME'

$ env TEXMFHOME="/tmp" kpsewhich --var-value TEXMFHOME

This is a bash shell

Do I do something wrong ? 

$ export TEXMFHOME="/tmp"
$ kpsewhich -var-value TEXMFHOME
$ kpsewhich -var-value '$TEXMFHOME'


Perhaps it is to be understood that if a texmf.cnf defines a variable then if this variable is queried, $ and ~
therein will be expanded ? 

I tried something else: 

$ export TMP1='$TMP2'
$ export TMP2='~'
$ kpsewhich -var-value '$TMP1'

$ kpsewhich -var-value '$TMP2'




jfbu | 30 Apr 14:51 2015

Speed of kpsewhich when querying the repertories of the TeX installation


upcoming TL2015 brings a great speed improvement to kpsewhich
on Mac OS X system, thanks to Adam Maxwell's analysis of the


but even after this great improvement, querying TEXMFHOME
and similar locations does still take some non vanishing time:

$ time kpsewhich --var-value=TEXMFHOME

real	0m0.090s
user	0m0.081s
sys	0m0.006s

This is on MacBook Air, and I get similar results on a i386 Linux box 

$ time kpsewhich --expand-path '$TEXMFHOME' 

real	0m0.109s
user	0m0.072s
sys	0m0.010s

A tenth of second just to query a repertory location seems
a lot to naive people unaware of the constraints (unknown to me) 

For example, if I want to write a package and provide
a way for the user to install it in a local or user tree, 
I may write a Makefile which will act after querying
the repertories locations. But one tenth of a second
for each kpsewhich call is a lot.

Could it be imagined to have some kpsewhichdir utility
which given TEXMFHOME will *instantly* return either
the environment variable $TEXMFHOME if it exists or the
TEXMFHOME setting as per the various texmf.cnf ?

best wishes