Re: PDF Files
> I would like to generate PDF files and they need to use type 1 fonts to be > readable. The times package supplies Type 1 fonts. > > Is the times package installed with teTeX, if not what do I have to do to > install it or is there another package I should be using for type 1 fonts? Get teTeX-1.0 if you have an older version. Then texdoc TETEXDOC and read section 2.7. Thomas
Re: help understanding fontimport
> I'm failing to understand something about how fontimport works. Well, the reason is that fontimport uses mktexnam (texmf/web2c/mktexnam) to find the "right" texmf tree for the font. The "right" texmf tree is the one where the sources (or other metrics) of the font can be found. If no such texmf tree can be determined, mktexnam says that the font should be stored in the varfonts area... > /var/tmp/texfonts/pk/epstypro/public/cbgreek/grmi1000.360pk Well, if you put the sources for cbgreek into a local texmf tree and list that texmf tree in texmf.cnf, then fontimport will work as you expect. Another (not so clean) solution is to set MT_DESTROOT to /usr/local/teTeX/texmf/fonts: env MT_DESTROOT=/usr/local/teTeX/texmf/fonts USE_VARTEXFONTS=0 fontimport -t `kpsexpand '$VARTEXFONTS'` I hope that this explains what happens... Thomas
Re: help understanding fontimport
> The mf sources for cbgreek fonts are in > /usr/local/teTeX/texmf/fonts/source/public/cbgreek > but the pkfonts are not copied to the right place. Hm... yes... I can think of a reason for this. mktexnam finds the tfm for the cbgreek fonts in the varfonts area and thinks that these fonts belong to that varfonts area (and not to the system's texmf tree). In mktexnam, we find: # Find the font: test tfm first, then mf, then possible sauterized mf. Maybe, the order should be reversed ... > > env MT_DESTROOT=/usr/local/teTeX/texmf/fonts USE_VARTEXFONTS=0 fontimport -t >`kpsexpand '$VARTEXFONTS'` > > That works, and reports the pk files will be moved into > .../texmf/fonts/..., not .../texmf.local/fonts... Well, this is brute force... Thomas
Re: use type 1 fonts
> I would like to use the type 1 fonts from Bluesky. After unpacking texmf I > see that those font are there. How do I tell if they are being used? Right > now, new pk fonts seem to be made every time I want to preview a document. Just texdoc TETEXDOC and read the sections about type1 fonts and pdf generation. > Also, after the upgrade, performance degraded. For example, with 0.4 the > texconfig utility would start after couple of seconds. Now it may take close > to a minute. Generation of fonts is also noticeably slower. Is this a > by-product of the size of the installation, or is there some sort of > optimization that I can do? Things are more flexible now and this costs some time. You can tune the scripts and replace kpsewhich calls by their results (i.e. hardcode the paths for directories and files) if you want. I think that the scripts can be made faster without loosing too much, but this is not a priority for me (at the moment). Thomas
Re: texconfig fails
> I'm trying very hard to install teTeX v1.0 on my Linux system. > Whenver I'm building the program, texconfig gives me this error: > > /usr/local/teTeX/bin/i686-pc-linux-gnu/texconfig: > /tmp/texconf6997/logfile: No such file or directory > > I see this line in the texconfig script: > mkdir $tmpdir || exit > But it's obviously not working right. What's wrong? texconfig detects a problem and wants to show the logfile, but obviously, the logfile does not yet exist. This is a bug in texconfig (which I think was fixed in teTeX-1.0.6), but not the main problem. You need to fix the main problem (which is what texconfig detects). I guess that you don't have the texmf tree in the right place. Did you follow the instructions given in the INSTALL file? Maybe, running texconfig init gives you a hint about what is wrong... Thomas
Re: File name Ambiguity in teTeX-1.0
> I just found a file name ambiguity in teTeX-1.0: > > >find /usr/local/apps/teTex/ -name 'verbatim.sty' -print > /usr/local/apps/teTex/share/texmf/tex/latex/tools/verbatim.sty > /usr/local/apps/teTex/share/texmf/tex/generic/misc2/verbatim.sty There is no "generic/misc2" directory in teTeX-1.0. > The same ambiguity has allready been in teTeX-0.4, but there the There is no "generic/misc2" directory in teTeX-0.4. > (correct) one from the tools package was included when running > LaTeX. In 1.0, the latter one is included. (How to configure this??) Just change the TEXINPUTS path (environment or texmf.cnf file). If you have an ambiguity between /path/to/a/b and /path/to/A/b, just put /path/to/a or /path/to/A before /path/to// in TEXINPUTS. Thomas
Re: texconfig fials
> Running mf to create plain base ... > This is METAFONT, Version 2.7182 (Web2C 7.3.1) (INIMF) > /opt/bin.link/texconfig[3]: 11190 Memory fault(coredump) > Done. > > Running mpost to create plain mem ... > This is MetaPost, Version 0.641 (Web2C 7.3.1) (INIMP) > /opt/bin.link/texconfig[3]: 11194 Memory fault(coredump) > Done. > > What could be the the problem ? Broken compiler: - core dump of mf and mp on HP-UX with -O2 + gcc 2.8.1. Recompiling mf1.c and mp1.c with -O seems to cure the problems. - gcc-2.95 and gcc-2.95.1 are known to reproduce wrong code on some platforms (e.g. sparc). If you have to use one of the above mentioned gcc versions, don't use -O2 optimization. -O should be ok. gcc-2.95.2 seems to work ok with -O2. If you want to use gcc, I recommend the latest version (which is gcc-2.95.2). Thomas
Re: texconfig fials
> One other thing I noticed, is strangs solvable make problem. > When you follow the quick install note for multi architecture platforms. > for HP ( I am using the Gnu make ); > ./configure > make world > then on Sun > make distclean > ./configure > make all > make install-exec > > This gives me a complete working set for the suns and for the HP we miss > some stuff like "xdvi.bin" but also some other things are not working > with texconfig. > But when I start again for the HP > make distclean > ./configure > make all > make install-exec > > Then the working set is compleet and all works. > So something in the make world goes wrong. "make distclean" should remove everything that configure and "make world" have generated in the source tree. Can you please check this? Thomas
Re: problem with mktextfm
> using teTeX 1.0 I have set the varibale $VARTEXFONTS > to collect the fonts there. This works fine for > mktexpk, but not for mktextfm. In the latter case > the .tfm as well as the .pk files end up in the current > directory. Why? This is a feature. If some user has his private metafont sources somewhere, .tfm (and .pk) files derived from these will not be put into a global texmf tree or the VARTEXFONTS area. Put the metafont sources into a system's texmf tree (listed in $SYSTEXMF) and the problem will hopefully disappear. Thomas
Re: problem with mktextfm
> you are right that this happens for private .mf files. > But in the case that I cannot put my .mf files in the > system tree as a ordinary user it would be nice to be > able to create the .tfm and .pk files in a certain > directory (not necessarily the working directory; may > be in the usual $VARTEXFONTS directory - it does not > really disturb other users, doesn't it). > How can I do this? You can set environment variables to override nearly everything. MT_DESTROOT the "root" of the directory tree, e.g. /some/texmf/fonts, but you can even set it to '$VARTEXFONTS' MT_PKDESTDIR destdir for pk fonts, e.g. '/my/pk/$MT_NAMEPART' Just look at the variables in texmf/web2c/mktexnam. Examples: setenv MT_DESTROOT '$VARTEXFONTS' # csh, tcsh, ... MT_DESTROOT='$VARTEXFONTS'; export MT_DESTROOT # sh, bash, ksh ... Thomas
Re: teTeX fuer NEXTSTEP wieder online
> ich habe meine Webseite ueber teTeX fuer NEXTSTEP auf unseren neuen Webserver > umgestellt, die Seite ist jetzt wieder unter der alten Adresse online. Thanks, Gregor. For those people on this list, who don't know what we are talking about: the web page about with lots of useful information about using teTeX on NEXTSTEP (thanks a lot, Gregor for maintaining this page!) is now (again) available at http://www.mathi.uni-heidelberg.de/~flight/stepTeX Thomas
Re: could some dvitty, dviselect and other similar tools be included in , tetex...
> or is there a better way to do these type of > tasks? Maybe... Someday... Not in the near future. Thomas
Re: Funny problem on teTeX beta 1.2.2000
> So why did the installation put the binaries on bin/i686-pc-linux-gnu, > if > teTeX cannot handle directories with such a name??? It works for me. How does your PATH look like? How does the TEXMFMAIN definition in texmf.cnf looks like? Can you please env KPATHSEA_DEBUG=4 kpsewhich -expand-var='$TEXMFMAIN' and check all files that are read? Thomas
Re: Problem compiling teTeX 1.0.7
> On compiling teTeX 1.0.7 on Sun Solaris 2.5 + gcc 2.95 That compiler is known to produce wrong code. Better avoid gcc 2.95 and gcc 2.95.1 (2.95.2 is ok). > ld: fatal: file /usr/local/lib/libXt.a: unknown type, unable to process using >elf(3E) libraries .. > I have checked the files libX11.a & libXt.a exist in /usr/local/lib Read the error message. The message does not say that libX11.a or libXt.a are missing. The linker just does not know how to handle them. So, these "libraries" seem to be unusable. Ask your sysadmin about them. If openwindows is installed, you can use the libraries in /usr/openwin/lib, but please "make distclean" and reconfigure. Thomas
Re: [Q] gzipped documentations
> In the debian distribution of tetex, all dvi files for packages are compressed > with gzip. It seems that the tools like texdoc, pksewhich and friends fail > becuase of that. Are there any plans to make them recognize compressed files ? Well, this feature can easily be added and I might do that some day. The real questions are: - are the debian people aware of the trouble caused by their compression (e.g. texdoc, doc/index.html won't work)? - are *they* going to fix these problems? Thomas
Re: [PATCH] texdoc and xdvi (Re: [Q] gzipped documentations)
> Content-Disposition: attachment; filename=texdoc.patch Interesting, but IMHO texdoc should do the decompression and not assume that the viewer does this. ghostview doesn't. acroread dosn't (compressing pdf does not make sense, but one can never be sure what people do). > Content-Disposition: attachment; filename=xdvi.patch Interesting, but one should not assume that the "tempfile" program is always available. One can prevent symlink attacks by creating a 700-mode directory and putting temp files there. This is portable, tempfile is not portable. If cleanup is done using event handlers (trap), then the event handler should already be set before the temp files are created... Thomas
Re: [PATCH] texdoc and xdvi (Re: [Q] gzipped documentations)
> > > Attached patch works with those lines in texmf.cnf:
> > > TEXDOCSSUFFIX = :.dvi:.ps:.html:.txt
> > > TEXDOCSCOMPRESS = :.gz:.bz2:.zip:.Z
> > > TEXDOCEXT = {$TEXDOCSSUFFIX}{$TEXDOCSCOMPRESS}
texmf.cnf should not be abused as general-purpose info system. Several
things which used to be in texmf.cnf for some time have been moved
to mktex.cnf. This was a good thing.
> Every (te)TeX distribution has its preferred documentation format.
The debian people should be careful not to break copyrights by omitting
.dvi and distributing .dvi.gz instead. koma-script (the new version as
distributed with version 1.0.1 of the texmf tree) for example requires
to distribute .dvi files.
> Yes, but it will imply some changes in current texdoc's behaviour, so
> Thomas will decide what is _the_ answer.
I have said that I'll propably add support for compressed doc files in
teTeX. I don't think that I have to give answers *now*. The debian people,
having decided to distribute a modified version of my texmf tree have to
find a solution for broken stuff (doc/index.html and texdoc).
texdoc is public domain, so people can do arbitrary changes.
BTW: Has anybody suggested decompression support in the xdvi *binary*
to Paul or Nicolai?
Thomas
Re: [PATCH] texdoc and xdvi (Re: [Q] gzipped documentations)
> modifications: on Win/UNIX, end of lines are different, so files > shipped on these platforms are different too. Does this break copyright? I don't consider "correcting" end-of-line markers a modification of a file. It is the correct representation of the file for a given OS. I don't know how a court would decide, however. > If not, why does compression break copyright? I did not say this. I said that the debian people have to be careful... Some package author might complain that his documentation becomes incorrect, because it has references to foo.dvi (and not to foo.dvi.gz). > Another most interesting point: some French people translate LaTeX > documentation. I currently do not understand how this is possible > without breaking the LPPL; I will some day ask on c.t.t. :) A lppl package lists all files that have to be distributed together. Adding files is usually ok (omitting files, however, usually not). > Come on. You always had useful comments on this thread, so i suggested > you to give us a hint before i send erroneous patches. A hint: don't execute viewers in the background. texdoc can clean up the temp directory after the viewer terminates: foo rm -rf ... Or, start a background job which starts the viewer and cleans up the temp dir after the viewer terminates: (foo; rm -rf ...) & Thomas
Re: (Fwd) Duplicate files?
> Did this ever get answered? If so I missed it. I did not answer yet. I am quite busy... > I just compiled teTeX-1.0.7 and noticed that it installs files under > $TEXMF/web2c, that are also installed by teTeX-texmf-1.0.1. The reason is that teTeX-texmf-1.0.1 must be complete enough to work with just the binaries. On the other hand, it is usually a good idea to install new pool files and stuff like this when building new binaries. In short: this is intentional. Thomas
Re: (Fwd) Duplicate files?
> places, I can see where you may want that, but that there might be a > difference between the two versions making it important which one was >From time to time (esp. before releases), I make a backup copy of the texmf tree and do a "make install" from the sources tree. Then, I check all differences of the two texmf trees and make sure that the files common in src / texmf are the same. Thomas
new stable and beta versions of teTeX
Hi, I have updated the stable version of teTeX: - the sources are updated from 1.0.6 to 1.0.7 - the texmf area is updated from 1.0 to 1.0.1 The 1.0.7 sources are mainly the same as the 1.0.6 sources; changes affect a few scripts and files that get installed into the texmf tree. The 1.0.1 texmf tree contains a lot of changes (most of these reflect updates of fonts / macros that happened since the initial 1.0 release). You can find the new stable 1.0.7 / 1.0.1 release at ftp://sunsite.informatik.rwth-aachen.de/pub/comp/tex/teTeX/1.0/distrib/ and soon on the mirrors (e.g. CTAN:systems/unix/teTeX/1.0/distrib). I have also uploaded a new version of the teTeX-beta release to the usual place: ftp://ftp.rrzn.uni-hannover.de/pub/local/misc/teTeX-beta/ (which is mirrored into CTAN:systems/unix/teTeX-beta). That beta version has a very similar texmf tree (but it is larger due to the czech/slovak support which is not available for the stable release). The sources in teTeX-beta reflect the current development state of most of the parts it contains (i.e. several parts are different from the stable release). Have fun! Thomas
Re: texmf.cnf
> we're using tetex 1.0.9 on a Solaris 2.6 OS. One of the users wishes to modify Hey, where did you get this? I don't know anything of a 1.0.9 version. > the value of the TEXMFCNF variable, to import his own texmf.cnf file. Set TEXMFCNF to the *directory* that the texmf.cnf file is in, not to the file itself. Thomas
Re: compilation problems
> Running mf to create plain base ... > Bus Error > Done. > > Running mpost to create plain mem ... > Bus Error > Done. This problem (gcc-2.95 and 2.95.1 on sparc-solaris) is mentioned in the PROBLEMS file in both: the latest stable and the latest beta version of teTeX. Thomas
Re: Panicing .. fonts messed up ..
> This is dvips(k) 5.86 Copyright 1999 Radical Eye Software (www.radicaleye.com) > ' TeX output 2000.02.27:2327' -> report.ps > kpathsea: Running mktexpk --mfmode ljfour --bdpi 600 --mag 1+57/600 --dpi 657 ph > vb8r I guess that dvips does not find the right psfonts.map file. Use find to locate all psfonts.map files on your disk. Use env KPATHSEA_DEBUG=4 dvips ... to see which of these files is used by dvips. The right file must have a line mentioning phvb8r. Thomas
Re: texmf.cnf for BLU
Hi Staszek et. al., > It is still not clear what can be changed in the local/user > configuration. For example, teTeX 1.0.7 texconfig do not make a copy of > texmf.cnf to the $VARTEXMF (if it was set, and as it was in previous > versions). This only makes sense, if you compile kpathsea with $VARTEXMF included in the search path for texmf.cnf files (before everything else). This is not the case in standard teTeX, so it is correct not to copy texmf.cnf into the VARTEXMF tree. It does not help to set TEXMFCNF inside some texmf.cnf file to something that includes $VARTEXMF. For texmf.cnf files, the *compile time* path is important (I think that we all agree on not to rely on environment variables). For TeX Live 5, it might be possible to do this change. One migth even consider to include $HOMETEXMF. > There are several reasons for having local|user|$VARTEXMF texmf.cnf: Sure, you are right. Thomas
Re: texmf.cnf for BLU
> The last one variable is rather not for ``where pdftex finds included > figures files!'', I'm rather sure that nobody put his figures here ;-), > but is _very important for finding first_ texmf/pdftex/config/psfonts.map > (which, in turn, can be different from texmf/dvips/config/psfonts.map). > > The last question: if pdf* do not use psfonts.map from the dvips/config/ > (or it use it as the last resort) why it use the same file name? Good questions. Playing games with search paths should be avoided whenever it is possible. > I remember Thomas' work with updmap. Perhaps old ideas are not so bad... :-) Thomas
Re: [BUG?] dvips with -Ppdf and mf-font
> >> > kpathsea: Running mktexpk --mfmode cx --bdpi 8000 --mag 1+0/8000 > >> > --dpi 8000 telefon ... Just don't use -Ppdf and better "just" use -Pwww. The 8000dpi "trick" to improve the quality of the pdf file does not work as long as bitmap fonts are involved (these will look bad in the resulting PDF output anyway). > Besides this, the reported message (and configuration) is an annoying behaviour > and IMHO not present in former versions. There was no config.pdf in former versions, so what are you complaining about? Thomas
Re: teTeX installation
> > /local/apps7/teTeX/bin/fmtutil: 11352 Memory fault: A memory image file is > > created as "core". > > fmtutil: `pdftex -ini -fmt=pdftex -progname=pdftex pdftex.ini' failed. > > OS: AIX 4.3.2 PPC > compiler: comes with the system, NOT the GNU compiler So, try to compile pdftex with less optimization. There are many buggy compilers around... Thomas
Re: texconfig -> Error opening terminal: generic
> When running texconfig, I get the following > texconfig -> Error opening terminal: generic > and the program exits with an exit status of 1. Is the texmf tarball installed correctly in $TEXMFMAIN? The terminfo entry sould be in texmf/texconfig/g/generic. You can debug texconfig by running it with sh -x `which texconfig` Thomas
Re: (pdf)latex.fmt not found
> I just completely erased my old teTeX installation and > reinstalled it. I think I have all the TEXINPUTS* > variables pointing to the right directories. Better don't sprcify the system directories for TEXINPUTS directly. It is too easy to do it wrong. A empty search component in TEXINPUTS soule better be used, as in: TEXINPUTS=:~/TeX// or TEXINPUTS=.:~/TeX//: or even: TEXINPUTS=.:~/TeX//::/some/other/dir > Is there anything I missed in the installation? Maybe. What does fmtutil --byfmt pdflatex say? And, does kpsewhich -progname=pdflatex pdflatex.ini find the pdftex/latex/config/pdflatex.ini file? Maybe, you have a messy TEXINPUTS path (which does not include texmf/pdftex)... The pdflatex format won't be build if kpsewhich cannot find pdflatex.ini. Thomas
Re: texmf.cnf for BLU
> TEXMFCNF = .:{$SELFAUTOLOC,$SELFAUTODIR,$SELFAUTOPARENT}\
> {,{/share,}/texmf{.local,}/web2c};c;/TeX/texmf/web2c
>
> If I changed that to
>
> TEXMFCNF = .:{$VARTEXMF,$SELFAUTOLOC,$SELFAUTODIR,$SELFAUTOPARENT}\
> {,{/share,}/texmf{.local,}/web2c};c;/TeX/texmf/web2c
>
> then the programs would search whatever was the value of VARTEXMF set
> in the texmf.cnf at compile time, yes? if someone changes VARTEXMF at
- $VARTEXMF should be $VARTEXMF/web2c
- variables on the right side of "=" will be expanded at runtime, not
at build time
- it does not work as expected :-(
To my last point: you usually set VARTEXMF in a texmf.cnf file and not
in the environment. But, for searching texmf.cnf files, only environment
variables and the compile-time path are considered, so there is no way
for a clean solution.
Sorry for suggesting using VARTEXMF in the search path of texmf.cnf.
It was a bad idea which does not work. I don't see an easy way (without
environment variables) which helps to maintain texmf.cnf in $VARTEXMF.
Thomas
Re: Xdvi and XFree86 4.0
[Cc: added for xdvi(k) maintainers] > I compiled and installed the new XFree86 4.0. It seems to work fine, but > teTeX's XDVI does not. It gives error message: > > Error: XtMakeGeometryRequest - parent not composite > > Do I have to recompile the teTeX, or what is the problem? I have no idea. Did you check the release notes of XFree86 4.0? Thomas
Re: on irix65 in batch mode the command `texconfig font ro' gives odd message
> tput: No value for $TERM and no -T specified If running in batch mode, try to set NO_CLEAR=true in the environment, e.g. as in env NO_CLEAR=true texconfig font ro Maybe, I should check with "tty -s" whether texconfig is connected to a terminal... Does anybody know of a UNIX system where tty -s && clear does not work as expected? Thomas
Re: env NO_CLEAR=true texconfig font ro doesn't work on irix65 in batch
> Subject: Re: env NO_CLEAR=true texconfig font ro doesn't work on irix65 in batch I see. There are a lot of instances of direct calls to the program clear. Can you replace them by cls and try again (except for the one call to clear in the body of the function cls itself)? > what is meant by tty -s && clear behaves as expected? what is If connected to a terminal: clear it. Otherwise: do nothing. Thomas
Re: Why is teTeX 1.0 much slower than the old 0.4
> One of our users noticed that our teTeX-1.0 is much slower than the old > teTeX-0.4. I guess that file searching is the cause. Please, compare the search paths (e.g. adding !! in the TEXMF definition in teTeX-1.0 might help a lot). Real debugging for file searching can be done by setting KPATHSEA_DEBUG (more about KPATHSEA_DEBUG is contained in the documentation for kpathsea). Thomas
Re: dvips abort
> dvips: xputenv.c:79: xputenv: Assertion `old_item' failed. > Aborted This problem has been analyzed and nothing wrong in the kpathsea code could be found. I guess that the problem is some incompatibility in the libc library. Does the libc6 binary that I provide work on both platforms? Thomas
Re: Installing fonts downloaded from CTAN
> gave nothing. Could you be more precise please? What is dvipsk to > begin with? $ dvips -version This is dvips(k) 5.86 Copyright 1999 Radical Eye Software (www.radicaleye.com) The program that you start with "dvips" is a dvipsk in fact. "dvips --version" gives an explicit statement about the kpathsea library being used. The manpage of dvips even says: ... ENVIRONMENT Dvipsk uses the same environment variables and algorithms for finding font files as TeX and its friends do. See the documentation for the Kpathsea library for details. (Repeating it here is too cumbersome.) ... > > the TDS documentation > I'll have a look. Just texdoc tds > A. "No, you don't need .afm, .pfb, .inf and .pfm" This is wrong. > B. "No, .fd, .tfm and .vf aren't sufficent. You need .pfb too." This is true. > If B, where to put the .pfb? TDS? texmf/fonts/type1/"supplier"/"typeface"/ Replace the strings "supplier" and "typeface" according to the font to be 100% correct. In fact, if you put your pfb files anywhere below texmf/fonts/type1/, it will be found. Thomas
Re: Fmtutil always return 0
> > > Consider the file test.sh
> > >#!/bin/sh
> > >trap 'rc=$?; echo $rc; exit $rc' 0 1 2 15
> > >exit 1
> > >
> > > And now, type
> > >sh test.sh || echo 1
> > >
> > > On HP-UX 10.20 with ksh, it prints 0.
Obviously, "exit 1" does not set $? to 1 on some systems. To fix the
matter for 0/1 exit codes, we could do some call to "test ..." or just
"false" to set the the return code to 1. A more general solution seems
to be the following:
exit_return() { setreturn "$@"; exit "$@"; }
setreturn() { return "$@"; }
I hope that this works on all systems (i.e. replace "exit 1" by
"exit_return 1") and add my two functions.
Can you please try:
#!/bin/sh
exit_return() { setreturn "$@"; exit "$@"; }
setreturn() { return "$@"; }
trap 'rc=$?; echo $rc; exit $rc' 0 1 2 15
exit_return 1
> > > On Linux with bash, 1 is printed twice.
The trap handler gets invoked twice. Which version of bash are you using?
You can turn tis off by writing the trap handler like this:
trap 'rc=$?; echo $rc; trap 0; exit $rc' 0 1 2 15
Thomas
Re: Fmtutil always return 0
> #!/bin/ksh > > changes behaviour. It doesn't echo number 1, but sets the exit code > correctly. Which means that the ksh does not execute the trap at all. Obviously, the bourne shell and the korn shell are not really compatible... Only at the very basic level... Thomas
Re: usage of stow
> Has it been considered to detect and then use stow by the installation > and texconfig to manage symlinks from the tetex tree to common places? > > http://www.gnu.org/software/stow/ No, but that should all work without any problem. teTeX, by default, already installs into a directory of its own. Symlinks are only needed for the binaries, manpages and info pages (which can be done by stow, if I understand that software correctly). Symlinks for the texmf stuff are not needed at all, because the selfauto* code can find the texmf tree relative to the (real) location of the binaries. Thomas
Re: Font Creation
> kpathsea: Running mktexpk --mfmode cx --bdpi 300 --mag 1+0/300 --dpi 300 > cmbx12 > mktexpk: Running hbf2gf -q -p cmbx12 300 > Couldn't open `cmbx.cfg' mktexpk runs hbf2gf with "test" option to check whether the font in question is supported by hbf2gf or not. If "hbf2gf" does not return an error code, mktexpk assumes that hbf2gf can handle that font and calls hbf2gf instead of metafont. So, to solve your problem, you have to check your hbf2gf configuration. Maybe, you just need a more recent version. Thomas
Re: fmtutil / kpathsea
>I have added the neccesary lines in fmtutil.cnf >latex209 tex - lplain.tex >slitex tex - splain.tex > >The problem is that these formats ask for >user input and at that point fmtutil aborts. >Is there an easy fix for this? What kind of user input is needed? Typing \dump? In that case, just create latex209.ini with just \input lplain \dump \endinput and use the line "latex209 tex - latex209.ini". I am not going to support user interaction in fmtutil. Thomas
Re: problem with xdvi
> sh: mktexpk: not found. ... > When I installed teTeX, I did not choose the option to create the symbolinc > links, but I do have symbolic links that are rather old and I suppose they Well, you can easily make the links "by hand". Example: cd /usr/local/bin rm -f `ls /usr/local/teTeX-beta-2201/bin/i686-pc-linux-gnu` ln -s /usr/local/teTeX-beta-2201/bin/i686-pc-linux-gnu/* . Thomas PS: Don't use /usr/local/teTeX-beta-2201/bin/i686-pc-linux-gnu/* in the second command. Big difference...
Re: multi platform support
> 1. what does it mean to have tetex installed as one tree? I think this means to install teTeX into a directory of its own. I.e. Not to install it directly into /usr/local where it shares /usr/local/bin whith other packgaes, but some directory like /usr/local/teTeX which is exclusively used by teTeX. > 2. to disable multi-platform support, is it enough for me to put the > binaries in teTeX/bin and not teTeX/platform/bin, or do i need to do If you install from sources: ./configure --disable-multiplatform If you use the precompiled binaries, use install.sh. It asks you whether or not to disable multi-platform support. Besides the other location for binaries, the only difference is the TEXMFMAIN line in the texmf.cnf file. Thomas
Re: xdvik doesn't find fonts
> % /usr/local/lib/texmf/dvips/config> cat config.pbv > p +pbv.map > > Running xdvi: > kpathsea: Running mktexpk --mfmode ljfour --bdpi 600 --mag 1+0/600 --dpi 600 pbvr8r > mktexpk: Running mf \mode:=ljfour; mag:=1+0/600; nonstopmode; input pbvr8r xdvi does not read pbv.map, because dvips -Pgsftopk does not either. It is documented, that gsftopk supports the same fonts as dvips -Pgsftopk. So, it should work if you add p +pbv.map to config.ps or config.gsftopk. Thomas
Re: xdvik doesn't find fonts
> Where? gsftopk(1) Thomas
Re: Powerpoint for tetex?
> But how, if I would like to show them using a beamer? I use pdflatex + pdfslides and do presentations using the acrobat reader. With http://www-sp.iti.informatik.tu-darmstadt.de/software/ppower4/, you can even get "fancy" effects, known from Powerpoint. Thomas
Re: tex API
> Ideally, I would like to have the (La)TeX interpreter available as a shared > library, which could then be invoked with something like: There is some code in web2c to make some functionality available via the IPC mechanism. This might or might not help you. I suggest you to ask your question on the [EMAIL PROTECTED] list, which is a more appropriate place. Thomas
Re: OT:Re: Powerpoint for tetex?
> Thomas: is there a particular reason for it not being included in > tetex? I did not find any offending copyright notice. The It did not plug in easily when I looked at it some time ago (problems including it into the source tree and problems with map files). I think that I'll add dvipdfm some time in the future. Thomas
Re: ttfont & tetex : help please
> My font "jvnr8a.pfb" has already install in : > /usr/share/texmf.local/fonts/pfb/microsft/verdana/ Change the search path for type1 fonts in texmf.cnf or rename /usr/share/texmf.local/fonts/pfb to /usr/share/texmf.local/fonts/type1 Then texhash again. Thomas
Re: pdflatex & pk font generation
> pdflatex -mktex=pk > > does the program do so. Is there an option in texconfig to set this to be the No. Thomas
Re: pdflatex & pk font generation
> Could it turn out to be a problem with other programs that use pdflatex if i > renamed the pdflatex binary (to something like "pdflatex.bin") and wrote a > shell wrapper that basically does something like "pdflatex.bin -mktex=pk $*" ? You can set MKTEXPK=1 in your environment or texmf.cnf. MKTEXPK.pdflatex (config file) or MKTEXPK_pdflatex (environment) is also possible. Renaming the executable is a bad idea. At least, you should try pdflatex.bin -mktex=pk -progname=pdflatex "$@" Thomas
Re: tetex & URW fonts
> In my tetex-1.0.6 (RH61) I can not found all necessary files for type1 > URW fonts (avantgar, bookman,courier,helvetic,ncntrsbk, palatino,times > ). Only pfb & afm are available. I installed tetex via RPM package. Is > this a error of RPM package or tetex distribution. Well, this is no error at all. I consider these URW fonts to be a free drop-in replacement of the LW35 Adobe fonts. Hence, you can just use the psnfss files + metrics to use these fonts, e.g. pslatex, times, etc. ... Look at the texmf/dvips/config directory. The map files (e.g. pdftex.map) use the URW fonts when the LW35 fonts are needed. I admit that it is confusing to use the metrics of ptmr* to use utmr8a.pfb, but this "trick" eliminates the need of a separate set of metrics + macros. > Anyway, I have to install these fonts by using fontinst :-) I don't think that this is necessary. Thomas
Re: acroread 4.0 generates PostScript `-infinity' from pdflatex output
> > Some users here are cautiously using pdflatex from teTeX 1.0.7 for the > > first time (pdflatex 3.14159-13d) > > is teTeX really only 13d? The "stable" teTeX release (1.0) basically is from june last year. The 7 patches did not update pdftex. > 3.14159-14f-released-2313, for what its worth. I am not sure I'd This is included in the latest teTeX beta. Thomas
Re: font problem
> When I use the letters 'fi' (like in benefit), I get them allright in the > DVI file, but get a pound sign in the PDF file instead. > Likewise I get a crossed circle sign for 'fl'. Sounds like some option tells dvips to enable the "character shifting" feature which is incompatible with the fonts you are using. > dvips $dvipsflag -Ptype1 -o $tempfile $infile Add -G0 (before -o) to the commandline of dvips. Thomas
Re: pdflatex and included pdf's
> work fine, but if I export an eps from xv, and run epstopdf on it and then > include the pdf, it doesn't render correctly in acroread, in spite of the I have made the same experience with pdftex versions before 0.14. If you use xv, better save as gif or png and include this into pdftex 0.13. Use the graphics package instead of the epsfig macros. This bug was fixed with pdftex 0.14, so your xv -> eps -> pdf route should work with pdftex 0.14. Thomas
Re: teTeX 1.0 vs. 0.4
> > Both epsf.sty and colordvi.sty are included in every *standard* TeX > > distribution, so something went wrong with your upgrade. Well, the files epsf.sty and colordvi.sty are not provided, but that should not be a problem, since the aliases files (in the main texmf tree) maps epsf.sty to epsf.tex and colordvi.sty to colordvi.tex. Have a *close* look at the following: $ kpsewhich colordvi.sty /usr/local/Text/teTeX-1.0/share/texmf/tex/plain/dvips/colordvi.tex $ kpsewhich epsf.sty /usr/local/Text/teTeX-1.0/share/texmf/tex/plain/dvips/epsf.tex $ egrep '^(colordvi|epsf)' /usr/local/Text/teTeX-1.0/share/texmf/aliases colordvi.tex colordvi.sty epsf.tex epsf.sty So, unless something is misconfigured, colordvi.sty and epsf.sty should be found by tex / latex. Thomas
Re: Can texconfig wipeout a harddisk?
Dear Ryan, > I am trying to find out if it is possible for the texconfig program > within teTeX 1.0.6 to wipe out a hard disk. Right now it is my Yes, I fear that this is true. Your mail is the second report of such a misbehaviour. George Tourlakis <[EMAIL PROTECTED]> has reported this three months ago (that was on a Nextstep system, too), but we could not find the reason for the misbehaviour. So, if anybody can help, I'd immediately fix texconfig. Thomas
Re: Can texconfig wipeout a harddisk?
Hi, here are my conclusions and a few answers to questions and comments about this subject. - using find / xargs xargs fails as soon as filenames contain whitespace (space or newline) - using find ... -print0 | xargs -0 That works fine with GNU tools, but I cannot assume that these are installed. - the current implementaion of fonts_ro / fonts_rw lacks a few quotes. This bug can cause wrong files being used as arguments for chmod, e.g. if you have the directory /some/texmf/fonts/pk/mode with space In this case, chmod will run on "/some/texmf/fonts/pk/mode" and on "with" and "space" (in the directory where texconfig is at that moment). Things like `...` in filenames do not cause command substitution (at least not here where no eval is used), so there is no reason to use grep on the find output. - using rm -f "$tmpdir"/*; rmdir "$tmpdir" is much better than rm -rf $tmpdir So, I think that I will use that. Especially, the missing quotes are a danger (imagine that someone has a space in $TMP, e.g. TMP='/ '). - absolute paths for basic commands are not portable - texconfig does not add `pwd` in front of $PATH, only to the end of $PATH (still, I do not remember why and I think that this can be removed) - As for empty globs in for i in *.log *.base; do rm -f "$bases/$i"; \mv "$i" "$bases/$i"; done they are not a big danger. Ok, you could lose the file with the name /some/dir/texmf/webc/*.log or /some/dir/texmf/webc/*.base but *not* all *.log or *.base files in the web2c directory. But, the code can be rewritten to for i in *.log *.base; do case $i in \*.log|\*.base) ;; *) rm -f "$bases/$i"; \mv "$i" "$bases/$i" ;; esac done So, there are three points for improvements (only one, however, really related to the original problem, because the other two improvements are minor pints). Thanks for all comments/suggestions. Thomas
(fwd) mktexlsr madness on NeXT
We are getting closer to the problem (look at the forwarded mail in the
end of my reply). I am sure that mktexlsr (alias texhash) was the reason
in the other two cases as well.
The only thing that fires after "normal" completion of mktexlsr are the
commands in the trap line. Can you please try the following script
(executed by /bin/sh) on your system:
#!/bin/sh
trap 'echo "-->$a<--"; exit 0' 0 1 2 3 7 13 15
a=hi
I guess that the output of this script is
--><--
instead of
-->hi<--
and that rm -rf behaves strange when called with "" as argument.
Maybe, you want to check this:
#!/bin/sh
mkdir $HOME/rmtest
touch $HOME/rmtest/somefile
cd $HOME/rmtest
rm -rf ""
find $HOME/rmtest -print
This should output the name of $HOME/rmtest and $HOME/rmtest/somefile.
Could you please check this?
Not to forget: many, many thanks for your efforts in anylyzing
this problem!
Thomas
---
>From [EMAIL PROTECTED] Mon Jun 19 00:51:05 2000
MIME-Version: 1.0 (NeXT Mail 3.3 v148.2.1)
From: HB2 <[EMAIL PROTECTED]>
Date: Sun, 18 Jun 2000 18:50:41 -0400
To: [EMAIL PROTECTED], [EMAIL PROTECTED]
Subject: texhash wiping out my harddisk
Gentlemen,
in the last few days I encountered the same problem reported by Ryan Scott.
Having reinstalled the basic software three (3) times now - I am getting
good at this - I am sure it has to do with teTeX 1.0.6.
My setup is the same, running Nextstep 3.3 on a NeXTstation plus the
NS33CISCUserPatch3 (to fix Y2K problems) on top of it. I downloaded the teTeX
files first from the primary site at then from DANTE. Installation went
smoothly including texconfig (fools luck?) but when I ran texhash it wiped
out my harddisk.
The third time I ran texhash from the "me"-account to prevent the root files
to be erased; this time it wiped out all the files in my account but at
least I could salvage some error messages from the console:
mktexlsr: /usr/local/teTeX/share/texmf: directory not writable. Skipping...
mktexlsr: Done.
rm: internal synchronization error: , ost+found, lost+found
rm: internal synchronization error: , rivate, private
rm: internal synchronization error: , cshrc, .cshrc
rm: internal synchronization error: , sr, usr
rm: internal synchronization error: , ev, dev
rm: internal synchronization error: , tc, etc
rm: internal synchronization error: , mp, tmp
rm: internal synchronization error: , ach, mach
rm: internal synchronization error: , dmach, sdmach
rm: internal synchronization error: , login, .login
rm: internal synchronization error: , profile, .profile
rm: internal synchronization error: , hidden, .hidden
rm: internal synchronization error: , path, .path
rm: internal synchronization error: , logout, .logout
rm: internal synchronization error: , NeXT, .NeXT
rm: internal synchronization error: , extLibrary, NextLibrary
rm: internal synchronization error: , extDeveloper, NextDeveloper
rm: internal synchronization error: , in, bin
rm: internal synchronization error: , dmach, odmach
rm: internal synchronization error: , extAdmin, NextAdmin
rm: internal synchronization error: , extApps, NextApps
rm: internal synchronization error: , e, me
rm: internal synchronization error: , oldcshrc, .oldcshrc
Apparently after mktexlsr finished its job it went on to erase all the files
in my "me" account but could not erase the basic mach, NeXT, bin, etc files
of the operating system - no permission as I ran texhash from "me". My path
was (and still is, I have all the TeTeX files, 100 MB of them, still sitting
in their default directory)
#
# this file gets executed by every shell you start
#
# make sure the path is correct
set path=(~/Unix/bin /usr/local/bin /usr/local/teTeX/bin /usr/ucb /bin
/usr/bin /usr/sybase/bin ~/Apps /LocalApps /NextApps /NextAdmin
/NextDeveloper/Demos .)
# setup umask to something reasonable
umask 022
# make the prompt palatable
if( ${?prompt} ) then
set host=`hostname`
set prompt="${host}> "
set history=50
# put aliases and other things down here
alias ls ls -F
alias mail /usr/ucb/Mail
endif
I might want to try to put /usr/localteTeX/bin first in the path but right
now I am pretty tired of reinstalling all my software and files; at least I
got another NeXTstation running teTeX 0.4 smoothly.
Any suggestions are most welcome,
Helmut Biritz
PS: Perhaps you might want to post this on the teTeX mail archive - right
now I am not even sure that my e-mail works properly as I had to reinstall
sendmail from scratch.
Re: dvi2tty maintainer
> I've made a few corrections and improvements to the dvi2tty package. > Whos's maintaining the software now? I need someone to pass my patches > The documentations seems too outdated to be trustable. I don't know > if this software has many users, but I'd like to be 100% compatible with GPL > in this subject. dvi2tty is unmaintained, I think. A successor is in development: CTAN:dviware/catdvi Antti-Juhani Kaijanaho writes: > I have uploaded catdvi-0.10.tar.gz into ftp.tex.ac.uk:/incoming. > The contents of that archive should replace the current content of > the directory CTAN:/dviware/catdvi/. > The catdvi program is licensed under the GNU General Public License, > version 2 or later. > > CatDVI is a DVI to plain text translator aiming to be a superior free > replacement for dvi2tty. In some areas it already outperforms dvi2tty, > while in some other areas it is clearly inferior to it. The program is > under development. Thomas
Re: howto update the texinfo source...
> I notice that the source packaged in tetex does NOT contain > the code for the info command which I would like to include This is intentional. I prefer to leave this program out of teTeX, because it has caused too much trouble in the past. E.g. a default compilation links agains shared libraries which are not available on other systems. If you want the info program, get the texinfo distribution and install that package according to its installation instructions. > into tetex. Perhaps the info source could be included in the next > distribution of tetex if it is not already being included. The I do not plan to do this. Thomas
Re: installing tetex
> However, fmt files are specific to a platform. If I want to mount the No. Just keep things as they are. > Where is getting that the bdpi (whatever that is) and resolution is 300. Maybe, you have some config.$PRINTER which says to use 300dpi. Or some ~/.dvipsrc. Use KPATHSEA_DEBUG=4 and see which files dvips opens. Thomas
Re: Where is `pdftex.map'?
> ..> kpsewhich pdftex.map > > gives no result. `pdftex' can find pdftex.map! Why not `kpsewhich'? kpsewhich wrongly guesses the file type: $ KPATHSEA_DEBUG=-1 kpsewhich pdftex.map 2>&1 | grep 'Search path for' | tail -1 kdebug:Search path for map files (from texmf.cnf) Let us try to find out what pdftex uses to find pdftex.map: $ echo '\relax 42\end' | KPATHSEA_DEBUG=-1 pdftex 2>&1 | egrep '(Search path for|pdftex.map)' | sed '/file=pdftex.map/q' | tail -2 | head -1 kdebug:Search path for PostScript header files (from texmf.cnf) Ok, now we know what to do: kpsewhich --format='PostScript header' pdftex.map Thomas
Re: \usepackage{times} and dvips -Ppdf
> If character shifting is needed it should be configured on a font-by-font > basis. Perhaps there should be a list of fonts to which shifting will be > applied when the option is set for dvips. This is really a dvips issue. The dvips maintainer knows about the problem that character shifting only works with *some*, but not with all fonts. So, we should just wait (or help him by sending a patch). Thomas
Re: texconfig and command line options
> It would be nice if texconfig had some command line options so that texconfig help Thomas
Re: texconfig and command line options
> Now how about adding this remark to the texconfig man page!!! Sure, that should be added to the manpage. Thomas
Re: Problem--inimp inimpost fail: Bus error
> I've recently built (a number of times) teTeX-1.0.7 (with update 001 and > update 002) on a new Solaris 2.8 system (with numerous patches) using GCC > 2.95.2 built on the same system. The configure and compilation step ... > Running mf to create plain base ... > Bus Error ... > Running mpost to create plain mem ... > Bus Error Exactly happended on earlier Solaris releases with gcc versions 2.95 and 2.95.1. The reason was an problem in the gcc optimizer. Compiling mf / mpost w/o optimization has helped. I guess that there is some bug in gcc 2.95.2 on Solaris 8, too. Workarounds: - compile with less optimization - use a different compiler - use binaries compiled by someone else Consider writing a bug report for gcc. Thomas
Re: texconfig and command line options
> Current setting: a4 Right, this is desirable. The same is true for the interactive part of texconfig. The directory names setup (for the mktex* scripts) section has this kind of feedback. I might implement more feedback of this kind, but not in the near future. So, if anybody wants to provide patches... Thomas
Re: TeTeX bug list?
> Is there somewhere a bug list (and maybe fixes) for tetex 1.0.7? Or there're no >known bugs > for 1.0.7? I maintain a non-public "TODO" list. Some items are bugfixes. Most reports / fixes are send through mailing lists (texk, tetex, tetex-pretest, pdftex), so reading them is a good idea... Thomas
Re: mathtime problem
> Cannot find font file mtsyn.pfb Mathtime fonts are commercial (get them e.g. from Y&Y). Alternatives: - mathptm(x) (uses free replacements from e.g. Symbol font; incomplete) - Belleek (attempts to be a free replacement for Mathtime; fonts have a few problems) Thomas
Re: ConTeXt formats
> Is there a reason why the context formats are commented out in the > fmtutil.cnf file? I'm sure that uncommenting it in the tetex Yes. I restrict teTeX to the formats which I consider "basic". These are plain and latex. I might change my view with respect of context so day... Thomas
Re: ConTeXt formats
Dear George, > 1) ConTeXt requires the table macro package, which AFAIK is not free, viz. > licen-en.pdf > > [3.5] This licence does not apply to components of the official version > that are provided by third parties. > > 2) The license says the sources are GPL's, but adds constraints on how you > can use ConTeXt, viz > > [5.3] Because the manuals provided by Pragma ADE are also examples of what > can be done by ConTeXt, their layout is bound to these documents. In > large this also applies to the whole ConTeXt demonstration suite. > Additional, third party documentation, therefore may not use the same > layout characteristics, like graphics (and tricks), unless permission is > granted. I know Hans very well and the change from the old very restrictive license to the better GPL was mostly triggered by discussions that I had with Hans. I am quite sure that if you tell Hans that his license has contradictions and unacceptable restrictions (from the free software point of view), he'll most propably change that. It seems like you understand the current Context licence better than me. Would you please be so kind to get in touch with Hans and discuss this with him? Thomas
Re: format file error..
> I am trying to integrate a more recent version of pdftex (14g) with my > teTex-1.0.7. Please, do fmtutil --byfmt pdflatex Do not forget to update hyperref, pdftex.def, supp-mis.tex, supp-pdf.tex available at http://tug.org/applications/pdftex/ Thomas
Re: LDAP-aware BibTeX [was: Barracuda, BibTeX and LDAP]
> My ldap2bibtex client's output for this entry is: ... > What I'd like to end up with is the integration of BibTeX with the > ldap searching procedure, such that each entry is resolved when > BibTeX is running instead of searching a static file. This is why Instead of hacking bibtex, you could write a tool (or extend your existing ldap2bibtex) to read LaTeX's aux files for citations and write a .bib file for just the documents that are cited. We are using such a tool (written in perl; Database is Oracle) for a long time now (> 10 years, earlier versions of this tool have been implemented in ESQL/C). Thomas
Re: Once more teTeX: xdvi stupidly runs ljfour
> some time ago, I posted the attached question, but I didn't get any
> answer. Is there really nonody wo can help?
This usually does not happen in teTeX, so I thought this was a debian problem...
> But xdvi stupidly makes fonts for mode ljfour. So, please, where do I
> find the REAL default mode for xdvi? And what does texconfig really do
In the xdvi wrapper script, replace the last line by
strac xdvstrace -e open xdvi.bin $NAMEOPT ${1+"$@"}
The strace output will show the names of all files which xdvi.bin opens.
Thomas
Re: PS-Fontdump
> I am looking for a program (linux/unix) that can dump (show/print) a > given pf(a/b)-Fontfile? http://gfontview.sourceforge.net/ Thomas
Re: Installation on AFS
> The files seen by the users are mounted on a > read-only volume /usr/local/vol/tetex/1.07 > but when I install it, I have to access the > read-write volume > /afs/blabla//usr/local/vol/tetex/1.07 So, just "tell" teTeX to use the read-write path for compilation. The search paths adjust on runtime, so do not care about the fact that the users see everything somewhere else. But, since kpathsea handles // specially, please avoid //. Change the default search path for TEXMFCNF is AFS fools the internal directory structure of the installation in a way that the texmf.cnf file cannot be found at runtime. Also, change the TEXMFMAIN definition in the installed texmf.cnf file if the automatic detection is wrong. Thomas
Re: teTeX-1.0.6
> that it was about the env-variable TEXMF, TEXINPUT. After we've removed > them it started working again. Why are they still there if not needed. Do not ask me why *you* set TEXMF and TEXINPUTS. *You* are responsible for your environment, not me. > Why can't tex tell me *where* it is searching for something, if it does kpsewhich -show-path=tex shows you the searchpath for tex (add -progname=latex for latex). > sorry for the harsh mail, but this wasn't neccessary... Everything you complain about is well documented... Thomas
Re: RPMs for Suse?
> On Thu, Sep 28, 2000 at 05:46:08PM +0200, Martin Schröder wrote: > > does anybody have up to date rpms for Suse Linux? It still ships > > with pdftex 0.13d :-( > > Hmmm ... we still ship teTeX-src-1.0.7 and teTeX-texmf-1.0.2, > is there anything newer _not_ beeing a beta release? There is nothing newer as far as teTeX is concerned. web2c-7.3.3 is released and this is compatible with pdftex-0.14f-released-2525 (look at CTAN:systems/web2c). So you could consider switching to your own TeX system build on top of web2c-7.3.3 (teTeX is still based on web2c-7.3.1) or "just" replace pdftex in teTeX (but, then, also update pdftex.def, hyperref, context, ...). I might do a major update of teTeX-beta (for web2c-7.3.3 + latest pdftex) in the near furure, but there won't be a new stable realease of teTeX in the next months. Thomas
Re: xdvi's buttons are missing
[Added xdvi + xdvik maintainers to Cc:] > I have "updated" my system to RedHat 7.0, and now my xdvi does not > have any of the buttons (Quit and friends). > > I'd like to emphasize that I have my own tetex installation, and I am > not using the RH rpms, so the only thing that changed is XFree and > bash to bash2 as the link to sh. I got > > $ rpm -q XFree86 bash > XFree86-4.0.1-1 > bash-2.04-11 > > Where should I look for clues? > > I have the output of `texconfig confall' as > > http://moni.msci.memphis.edu/~mw/texconfig_confall.txt I can confirm that. It even happens for me if I recompile xdvi. Running xdvi against the XFree-4.01 shared libs shows this problem. Running the same binary arainst the XFree-3.?? libraries works ok. Thomas
Re: xdvi's buttons are missing
> Have you tried -- > plain (non-k) xdvi? I just did (for xdvi-22.29): ./configure --with-tetex make TEXMFMAIN=/usr/local/teTeX/share/texmf TEXMFCNF=/usr/local/teTeX/share/texmf/web2c ./xdvi /tmp/small2e and the buttens were not missing. One warning (error?) was displayed, though: ./xdvint_line: No such file or directory > xdvi +expert > the 'x' keystroke No buttons with xdvik. Or, more precise, the place for the buttinlist is there, but it is empty. Thomas
xdvi-22.29
Hi, first, a further analysis of messages displayed by xdvi-22.29; an analysis of the "button problem" follows later. xdvi-22.29 error messages on startup: Starting xdvi-22.29 from the build directory with TEXMFMAIN=/usr/local/teTeX/share/texmf TEXMFCNF=/usr/local/teTeX/share/texmf/web2c ./xdvi /tmp/small2e shows the ./xdvint_line: No such file or directory error when xdvi is compiled/linked against XFree86-3.X, too. The message vanishes if I install xdvi properly, i.e. copy the binary to the install-directory and set PATH and invoke it by "xdvi ...". If I invoke xdvi by its full path, i.e. /usr/local/teTeX/bin/xdvi/xdvi /tmp/small2e I get the following error message: /usr/local/teTeX/bin/xdvi/xdvietaPost as well.: No such file or directory The program is working fine, though (despite of the error message). So, I guess that the messages depend more on the way xdvi is invoked and not on the X version. button problem: I have had a closer look at the button problem of xdvik and concluded the following: - the problem is caused by compiling against the XFree86-3.X headers and running against the shared libs of XFree86-4.01 - there is no difference between plain xdvi or xdvik regarding this bug (even though I have claimed that only xdvi works ok with XFree86-4.01 in an earlier message. That was wrong.) Thomas
Re: version
> What is the difference between 1.0.7 and 1.0.6? I can't seem to find a > detailed changelog anywhere. Just some cleanups in scripts. Binaries are the same (in 1.0.6 and 1.0.7). $ diff -ru a b | grep '^diff' diff -ru a/teTeX-1.0/PROBLEMS b/teTeX-1.0/PROBLEMS diff -ru a/teTeX-1.0/texk/gsftopk/render.ps b/teTeX-1.0/texk/gsftopk/render.ps diff -ru a/teTeX-1.0/texk/kpathsea/mktexlsr b/teTeX-1.0/texk/kpathsea/mktexlsr diff -ru a/teTeX-1.0/texk/tetex/dvired b/teTeX-1.0/texk/tetex/dvired diff -ru a/teTeX-1.0/texk/tetex/epstopdf b/teTeX-1.0/texk/tetex/epstopdf diff -ru a/teTeX-1.0/texk/tetex/fontimport b/teTeX-1.0/texk/tetex/fontimport diff -ru a/teTeX-1.0/texk/tetex/texconfig b/teTeX-1.0/texk/tetex/texconfig diff -ru a/teTeX-1.0/texk/tetex/texexec b/teTeX-1.0/texk/tetex/texexec diff -ru a/teTeX-1.0/texk/tetex/texlinks b/teTeX-1.0/texk/tetex/texlinks diff -ru a/teTeX-1.0/texk/tetex/texutil b/teTeX-1.0/texk/tetex/texutil diff -ru a/teTeX-1.0/texk/tetex/thumbpdf b/teTeX-1.0/texk/tetex/thumbpdf diff -ru a/teTeX-1.0/texk/web2c/fmtutil.in b/teTeX-1.0/texk/web2c/fmtutil.in diff -ru a/teTeX-1.0/texk/web2c/share/bitstrea.map b/teTeX-1.0/texk/web2c/share/bitstrea.map diff -ru a/teTeX-1.0/texk/web2c/share/monotype.map b/teTeX-1.0/texk/web2c/share/monotype.map diff -ru a/teTeX-1.0/texk/web2c/share/variant.map b/teTeX-1.0/texk/web2c/share/variant.map Thomas
Re: Rebuilding teTeX format files?
> main_memory.context = 110 > main_memory = 263000 % words of inimemory available; also applies to inimf&mp > > ). What is `.context'? Is there a (required) relationship between `main. > memory.context' and `main.memory'? Incidentally, why such odd numeric values? Ignore the .context values. context is a macro package (such as latex). If you do not know what context is, you are propably not using it... > I gather that I have to rebuild all the format files of the installation > to apply a change. Is there a script to do this? Failing that, which texocnfig init or fmtutil --all Thomas
Re: Error opening terminal: generic.
> If you look at the texconfig, it uses the dialog program > (atleast on Linux and BSD). So, it seems you have not defined > your TERM environment variable. Try xterm or vt100, depending on > what kind of terminal (emulator) you use. texconfig ignores the environment variable $TERM (and sets TERM=generic for the dialog call) if it decides to use its own dialog program. That only works, however, if the terminfo file $TEXMFMAIN/texconfig/g/generic is in its place. On a FreeBSD or Linux system, texconfig uses the system's /usr/bin/dialog (if it exists) and does not use its own "generic" terminal settings. Thomas
licenses in teTeX
Dear TeX user, I have maintained teTeX for over 5 years and now I need your help in order to make it really "free software". The first step is to find all files which are not covered by a free software license. Any file covered by a license not mentioned in the list given below should be reported to [EMAIL PROTECTED] (please, do not send such reports directly to me). If you want to help, make sure to have a teTeX-beta release which is dated from 4 Aug 2000 or later. Please report any file which does not have one of these licenses: - LaTeX Project Public License (LPPL, see http://www.ctan.org/tex-archive/macros/latex/base/lppl.txt) - GPL (see http://www.gnu.org/copyleft/gpl.html) - LGPL (see http://www.gnu.org/copyleft/lgpl.html) - Public Domain (PD, see http://www.gnu.org/philosophy/categories.html#PublicDomainSoftware) - the old BSD license or the revised BSD license Other licenses might or might not be ok, so we want to find them and check them. If you want to help, please make sure you have the *original* teTeX-beta version released by me. The archives which need license checking are: ftp://ftp.dante.de/pub/tex/systems/unix/teTeX-beta/teTeX-texmf-beta-2804.tar.gz and ftp://ftp.dante.de/pub/tex/systems/unix/teTeX-beta/teTeX-texmfsrc-beta-2804.tar. gz You can find these files on other CTAN servers, too, so please use a mirror near to you. BTW: for the third teTeX tarball (teTeX-src-beta-2804.tar.gz, containing the sources for the binaries), we are quite sure that every file is covered by a free software license. Thank you for your help! Thomas Esser
Re: ec - fonts
> What did I miss? To write a proper bug report which enables other people to reproduce your problem? Thomas
Re: dvips can't find font file phvr8r.tfm
> $ dvips -o presentation.ps presentation.dvi > This is dvips(k) 5.86 Copyright 1999 Radical Eye Software (www.radicaleye.com) > ' TeX output 2000.10.23:1627' -> presentation.ps > dvips: Can't open font metric file phvr8r.tfm > dvips: I will use cmr10.tfm instead, so expect bad output. > dvips: ! I can't find cmr10.tfm; please reinstall me with proper paths Something has messed up dvips's search paths for tfm and vf files. Check your conifg file + environment variables. If it all does not help, set KPATHSEA_DEBUG=-1 (environment) and rerun dvips. Read the output carefully. If that does not give you the clou, send the output to me (not to the list). Thomas
Re: ec - fonts
> So, the ec-fonts are not installed on teTeX-beta-2807 ? The simple "driver files" are generated "on the fly". Just try to use, e.g. ecrm1000. Thomas
Re: teTeX betas
> I see mention on this list of `teTeX-beta-2807', yet when I look > under the link `beta versions' on `The teTeX Homepage' (http://www.tug.org/ > tetex/) the FTP directory there is empty. > > Shouldn't the directory have something in it? The admin of ftp.rrzn.uni-hannover.de says that they have problems to mirror wuftp servers version 2.6.0 or later. Can anybody help them? If so, please get in touch with [EMAIL PROTECTED] Thomas
Re: Metafont: X display under Red Hat Linux
> 29) Metafont does not support the X display. > It may be that your Metafont binary is compiled without support for > the X windows system. If there exists a mf.X binary, that one > has the missing X support you need. This is a bug in the docs. The name of the program is mfw. Sorry... Thomas
Re: problems with bookman.sty
> $ xdvi a& > $ kpathsea: Running mktexpk --mfmode ljfivemp --bdpi 600 --mag 1+0/600 > --dpi 600 pbkl7t > mktexpk: Running mf \mode:=ljfivemp; mag:=1+0/600; nonstopmode; input > pbkl7t The problem is that xdvi does not find pbkl7t.vf. Does some environment variable force xdvi to use a wrong search path? Use KPATHSEA_DEBUG to find out more about this problem if your environment is ok... Thomas
Re: Putting texmf.cnf in a non-standard place
> I can do this after the fact by setting the TEXMFCNF environment variable > to point to the directory containing the preferred version texmf.cnf -- is Change TEXMFCNF in texk/kpathsea/texmf.in to whatever you need. Thomas
Re: still problems with bookman
> [...] > > kpathsea: Running mktexpk --mfmode ljfivemp --bdpi 600 --mag 1+0/600 > > --dpi 600 pbkl7t > ... > You seem to be missing the font outlines. xdvi needs them to > create the preview. Just create ps file and preview it with gs. This interpretation misses the point: the problem here is that xdvi does not find pbkl7t.vf. That file tells xdvi to look for pbkl8r... Looking at the strace output of the xdvi run, I have noticed that xdvi does not read the ls-R file of the main texmf tree. I think that either - this important file does not exist or - the search path for ls-R files is wrong The !! that is specified in the search path for vf fonts disallows xdvi to search the directories. Thomas
Re: now I think this is a bug
> export PATH=/usr/local/teTeX/bin/i686-pc-linux-gnu:$PATH ... > TEXMFMAIN ``/usr/local/teTeX/bin/i686-pc/share/texmf'' seems to have Whow. Cool. Can you please $ kpsewhich -expand-var='$SELFAUTOLOC -- $SELFAUTODIR -- $SELFAUTOPARENT' I, for example, get the following output: /x/t/i/teTeX/bin/i586-pc-linux-gnu -- /x/t/i/teTeX/bin -- /x/t/i/teTeX You should get /usr/local/teTeX/bin/i686-pc-linux-gnu -- /usr/local/teTeX/bin -- /usr/local/teTeX > # rpm -q glibc > glibc-2.2-5 > > I'll be glad to debug this---if you give me some pointers. This glibc is a very recent version. So, the rest of your system is likely to be very "bleeding edge", too? Well, I suspect that we run into some compiler bug here... Thomas
Re: now I think this is a bug
> # kpsewhich -expand-var='$SELFAUTOLOC -- $SELFAUTODIR -- $SELFAUTOPARENT' > /usr/local/teTeX/bin/i686-pc-linux-gnu -- /usr/local/teTeX/bin/i686-pc-linux -- >/usr/local/teTeX/bin/i686-pc Looks very broken. This sould be impossible. Look at the kpathsea code (progname.c, I think). Must be a compiler bug. > This glibc is the official update of RH to their 7.0 distro. I have > absolutely nothing on my system which is not official RH (except tetex). ... > # gcc -v > Reading specs from /usr/lib/gcc-lib/i386-redhat-linux/2.96/specs > gcc version 2.96 2731 (Red Hat Linux 7.0) And still: I think that your compiler is broken. Thomas
Re: suggestion: `EPM'
> EPM is a free UNIX software/file packaging program that generates > distribution archives from a list of files. The distributions > include installation and removal scripts that handle such details > as diskless client installations and initialization scripts. > > EPM can also generate "native" software distributions for Debian, > HP-UX, IRIX, Red Hat Linux, and Solaris. > > ==> http://www.easysw.com/epm/ Well, I need one package format which can be used easily on *all* platforms that I support. I do not want to provide different "native" software distributions for teTeX. This is what the UNIX distributions can do, e.g. FreeBSD, Irix, Solaris, ... Even commercial UNIX vendors provide teTeX packed in "their" native package format (in their "freeware" area), so I do not see any need that I change something. Thomas
Re: Error in top level configure script
> In teTeX-1.0.7, in the top level configure script, the variable > needs_pnglib is carefully set, but on line 2539, needs_libpng is > referred to instead of needs_pnglib. Thanks. The fix is to replace needs_libpng by needs_pnglib in libs/libpng/libpng.ac and to run the reautoconf script (at top-level). Thomas
Re: texdoc + Netscape
> I just discovered that when texdoc finds an .html file, it
> tries to start a new Netscape, using the command
Right. I have changed texdoc a few months ago in my sources:
: ${TEXDOCVIEW_html='netscape -remote openURL'"'(%s)'"' 2>/dev/null || netscape %s
&'}
Anyway, thanks for your suggestion.
Thomas
