A tarball candidate is at http://www.pdftex.de/tex/pdftex/pdftex-1.20a-rc1.tar.bz2 Best regards Martin -- Martin Schröder, ms@artcom-gmbh.de ArtCom GmbH, Lise-Meitner-Str 5, 28359 Bremen, Germany Voice +49 421 20419-44 / Fax +49 421 20419-10 http://www.artcom-gmbh.de
What is the difference between pdfinitex and pdftex --ini? hahe@hahepc1:~$ pdfinitex y This is pdfeTeXk, Version 3.141592-1.20a-rc1-2.1 (Web2C 7.5.2) %&-line parsing enabled. kpathsea: Running mktexfmt pdfinitex.efmt fmtutil: no info for format `pdfinitex'. I can't find the format file `pdfinitex.efmt'! hahe@hahepc1:~$ pdftex --ini y This is pdfeTeXk, Version 3.141592-1.20a-rc1-2.1 (Web2C 7.5.2) (INITEX) %&-line parsing enabled. output format initialized to DVI (./y.tex ! Missing { inserted. <to be read again> { l.1 \hbox{ abc} ? Shouldn't the first behave as the second? Man page says --ini: Be pdfinitex... Regards, Hartmut
problem around texmfmp.c: inipdftex call gives in main.c of texmfmp.c (= pdfetexextra.c), with some debug printf() lines: =========argv[0] <inipdftex> =========kpse_program_name <inipdftex> =========INI_PROGRAM <pdfeinitex> so it does not match the INI_PROGRAM, therefore it doesn't go into ini mode. inipdfetex doesn't either, obviously. But when I do ln -s pdfetex pdfeinitex pdfeinitex call gives: hahe@hahepc1:~$ pdfeinitex y =========argv[0] <pdfeinitex> =========kpse_program_name <pdfeinitex> =========INI_PROGRAM <pdfeinitex> This is pdfeTeXk, Version 3.141592-1.20a-rc1-2.1 (Web2C 7.5.2) (INITEX) %&-line parsing enabled. So somehow the INI_PROGRAM name matching would need an extension so that more combinations for INI_PROGRAM name are possible. Particularily the names starting with inipdf... currently have no use, it seems. Regards, Hartmut
To match more ini-names, here is a quick'n dirty kludge in texmfmp.c: --- /usr/local/src/tex/pdftex-1.20a-rc1/src/texk/web2c/lib/texmfmp.c Mon Mar 8 22:47:26 2004 +++ texmfmp.c Sat Mar 20 23:38:48 2004 @@ -200,6 +208,18 @@ if (readyalready != 314159) { /* The `ini_version' variable is declared/used in the change files. */ boolean virversion = false; +#if defined(pdfTeX) || defined(pdfeTeX) + int i; + char *pdf_ini_prognames[] = INI_PROGRAMS; + char *pdf_vir_prognames[] = VIR_PROGRAMS; + for(i = 0; *pdf_ini_prognames[i] != 0; i++) + if (FILESTRCASEEQ (kpse_program_name, pdf_ini_prognames[i])) + iniversion = true; + for(i = 0; *pdf_vir_prognames[i] != 0; i++) + if (FILESTRCASEEQ (kpse_program_name, pdf_vir_prognames[i])) + virversion = true; + /* this is once again checked below, don't mind */ +#endif if (FILESTRCASEEQ (kpse_program_name, INI_PROGRAM)) { iniversion = true; } else if (FILESTRCASEEQ (kpse_program_name, VIR_PROGRAM)) { and in both pdftexextra.h and pdfetexextra.h I have added: #define INI_PROGRAMS {"pdfinitex", "pdfinietex", "pdfeinitex", "inipdftex", "inipdfetex", ""} #define VIR_PROGRAMS {"pdfvirtex", "pdfviretex", "pdfevirtex", "virpdftex", "virpdfetex", ""} Then it seems to get into ini-mode when called by the right name. Or, more simple, just search for a ini/vir substring in the name? Shouldn't one add something like this? Regards, Hartmut
Shouldn't one add something like this?
I don't think so. This mess with ini- and vir- stuff is non-sense, look at what you found out as an evidence. I think we should remove every single *ini* and *vir* name from the list of binaries. Fabrice
On Sun, 21 Mar 2004, Fabrice Popineau wrote:
Shouldn't one add something like this?
I don't think so. This mess with ini- and vir- stuff is non-sense, look at what you found out as an evidence.
It _is_ a mess.
I think we should remove every single *ini* and *vir* name from the list of binaries.
Yes, at least the ones not working should be removed, and then the rest, for consistency :-) But what about INITEX, it's in the TeXbook? Regards, Hartmut
Yes, at least the ones not working should be removed, and then the rest, for consistency :-) But what about INITEX, it's in the TeXbook?
web2c has the programs "TeX" and "INITEX" in one binary. There is nothing wrong with this, I think. Thomas
I think we should remove every single *ini* and *vir* name from the list of binaries.
Yes. TeX Live, teTeX and web2c go into this direction, so putting any work into this *ini* and *vir* names makes no sense. People can either use tools (like fmtutil) or cmd line options (like -ini). Thomas
At 00:01 21/03/2004, Fabrice Popineau wrote:
Shouldn't one add something like this?
I don't think so. This mess with ini- and vir- stuff is non-sense, look at what you found out as an evidence.
I think we should remove every single *ini* and *vir* name from the list of binaries.
indeed, no need for that any more, actually, for me they are already non existent, --ini is enough! Hans
At 20:18 20/03/2004, you wrote:
What is the difference between pdfinitex and pdftex --ini?
hahe@hahepc1:~$ pdfinitex y This is pdfeTeXk, Version 3.141592-1.20a-rc1-2.1 (Web2C 7.5.2) %&-line parsing enabled. kpathsea: Running mktexfmt pdfinitex.efmt fmtutil: no info for format `pdfinitex'. I can't find the format file `pdfinitex.efmt'!
hahe@hahepc1:~$ pdftex --ini y This is pdfeTeXk, Version 3.141592-1.20a-rc1-2.1 (Web2C 7.5.2) (INITEX) %&-line parsing enabled. output format initialized to DVI (./y.tex ! Missing { inserted. <to be read again> { l.1 \hbox{ abc} ?
Shouldn't the first behave as the second?
i'd think so, maybe some progname confusion, i.e. they have different 'names' unless (in windows:) we would have: pdfinitex.bat contains: @echo off texmfstart texexec.pl %* Hans
participants (5)
-
Fabrice Popineau -
Hans Hagen -
Hartmut Henkel -
Martin Schröder -
Thomas Esser