new context and luatex packages for Debian
Following Taco's release of the new luatex here is now the announcement
of new packages for context and luatex.
Both have been uploaded to Debian/experimental, but can be obtained
directly via:
deb http://people.debian.org/~preining/TeX/ context/
deb http://people.debian.org/~preining/TeX/ luatex-sid/
Version: luatex 0.30.3, context 2008.10.31
Any comments are appreciated.
I hope to get the MarkIV stuff sorted out sooner or later...
Best wishes
Norbert
-------------------------------------------------------------------------------
Dr. Norbert Preining
Following Taco's release of the new luatex here is now the announcement of new packages for context and luatex.
Good news, thanks!
Both have been uploaded to Debian/experimental Version: luatex 0.30.3, context 2008.10.31
CONTEXT ======= The new context passed my usual tests: compiling a couple of 100+-page books with math and physics equations and figures. It didn't pass a new test, mixing Palatino and Euler: ========== start test.tex =============== \usetypescript[palatino] [ec] \definetypeface [palatino] [mm] [math] [euler] [euler] [rscale=1.03] \setupbodyfont[palatino,11pt] \starttext \startformula c^2 = a^2 + b^2. \stopformula \stoptext ========== end test.tex =============== This test also did not work with my previous context installation (the 2008.05.21 packages). With either the current or previous packages, the equation comes out in Euler but as "a^2 b^2 + c^2:" (i.e. missing equals sign, and the period replaced by a colon). The test works on live.contextgarden.net (context 2008.11.10), and it worked with an older set of context minimals a month or so ago. So I suspect a problem in my font map files (I'm using lmodern 1.010x-5). The problem might be caused by something silly that I did; however, as far as I know I run a vanilla tex setup from your Debian 'unstable' packages. Let me know if you have the same problem. LUATEX ====== I couldn't find luatex in Debian/experimental. I tried two methods to find it. The first method is: $ aptitude update # with experimental in my sources.list $ aptitude install luatex/experimental It produced: Unable to find an archive "experimental" for the package "luatex" Which was strange, since I also have your line deb http://people.debian.org/~preining/TeX/ luatex-sid/ in my sources.list, and it wasn't picked up there either. For the second method, I visited http://packages.debian.org/search?suite=experimental&keywords=luatex which lists these luatex packages * experimental (tex): next generation TeX engine 0.30.3-1: amd64 hppa ia64 kfreebsd-amd64 kfreebsd-i386 sparc So, there's no i386 package. I'm using a TP T60, so that's why the first method didn't find the new luatex. I've just downloaded the source package and, as a check, will try to build the binary package. -Sanjoy `Until lions have their historians, tales of the hunt shall always glorify the hunters.' --African Proverb
Sanjoy Mahajan wrote:
========== start test.tex =============== \usetypescript[palatino] [ec] \definetypeface [palatino] [mm] [math] [euler] [euler] [rscale=1.03] \setupbodyfont[palatino,11pt]
\starttext \startformula c^2 = a^2 + b^2. \stopformula \stoptext ========== end test.tex ===============
This test also did not work with my previous context installation (the 2008.05.21 packages). With either the current or previous packages, the equation comes out in Euler but as "a^2 b^2 + c^2:" (i.e. missing equals sign, and the period replaced by a colon).
Works for me (but then, I am not using a 100% typical minimal). It looks like a problem with a virtual font. This in turn is often caused by an extra or wrong map line, but it could also be a missing .vf. Is there no hint in the log? Best wishes, Taco
Hi Sanjoy, thanks for testing. On So, 30 Nov 2008, Sanjoy Mahajan wrote:
The new context passed my usual tests: compiling a couple of 100+-page books with math and physics equations and figures. It didn't pass a new
That is *really* great. In fact, as stated, if you don't mind me having it on my computer (but I guarantee that I will not use it for anything but testing context) I would be happy to have a big fat test document. You already sent me one, that is one of those I test before uploading.
========== start test.tex =============== \usetypescript[palatino] [ec] \definetypeface [palatino] [mm] [math] [euler] [euler] [rscale=1.03] \setupbodyfont[palatino,11pt]
\starttext \startformula c^2 = a^2 + b^2. \stopformula \stoptext ========== end test.tex ===============
This test also did not work with my previous context installation (the
Strange. I am running it here on my laptop and it comes out fine, Debian packages: context 2008.10.31-1 lmodern 1.010x-5
worked with an older set of context minimals a month or so ago. So I suspect a problem in my font map files (I'm using lmodern 1.010x-5).
Pity that -recorder cannot be passed to texexec so that it records *all* loaded files. This is the best way to find interfering files. Maybe texexec has something similar?
I couldn't find luatex in Debian/experimental. I tried two methods to
It hasn't been compiled for x86 till now, I uploaded amd64 binaries.
Thanks a lot for all your tests, much appreciated.
Best wishes
Norbert
-------------------------------------------------------------------------------
Dr. Norbert Preining
On Mo, 01 Dez 2008, Taco Hoekwater wrote:
texexec --passon='-recorder' (but it doesn't list ruby files)
Great to hear, thanks.
Best wishes
Norbert
-------------------------------------------------------------------------------
Dr. Norbert Preining
In fact, as stated, if you don't mind me having it on my computer (but I guarantee that I will not use it for anything but testing context) I would be happy to have a big fat test document. You already sent me one, that is one of those I test before uploading.
One of my two tests is a slightly newer version of the mathematics textbook that I sent you. I can send you the newer math textbook if that's useful. Now that the Euler problem is fixed (see below), I'll probably use Euler for the equations, so you'll have a test suite using Palatino and Euler. The other test is the physics textbook ("Art of Approximation"). The physics textbook uses two Python scripts that I wrote plus mercurial. If you don't mind testing Python as a side effect of testing the Debian packages, I can send you a version of the physics textbook that still uses Python but doesn't depend on mercurial (or I can send you the vanilla version with all those deps).
Strange. I am running it here on my laptop and it comes out fine, Debian packages: context 2008.10.31-1 lmodern 1.010x-5
worked with an older set of context minimals a month or so ago. So I suspect a problem in my font map files (I'm using lmodern 1.010x-5).
Pity that -recorder cannot be passed to texexec so that it records *all* loaded files. This is the best way to find interfering files.
that suggestion solved the problem. You can pass -recorder as follows: texexec --passon=-recorder test.tex pdftex then gets the -recorder option, and test.fls will have the magic info. I did so and looked for any loaded files that begin with /home/sanjoy. And found the culprit: INPUT /home/sanjoy/texmf/tex/context/math-eul.tex So I had an interfering version of math-eul.tex. When I deleted it, the test file compiled fine. Sorry for the noise. -Sanjoy `Until lions have their historians, tales of the hunt shall always glorify the hunters.' --African Proverb
Hi Sanjoy, On Mo, 01 Dez 2008, Sanjoy Mahajan wrote:
One of my two tests is a slightly newer version of the mathematics textbook that I sent you. I can send you the newer math textbook if that's useful. Now that the Euler problem is fixed (see below), I'll probably use Euler for the equations, so you'll have a test suite using Palatino and Euler.
Great!!!
The other test is the physics textbook ("Art of Approximation"). The physics textbook uses two Python scripts that I wrote plus mercurial. If you don't mind testing Python as a side effect of testing the Debian packages, I can send you a version of the physics textbook that still uses Python but doesn't depend on mercurial (or I can send you the vanilla version with all those deps).
I have mercurial 1.0.1-5.1 installed, but I guess you were talking about access to the mercurial repository you are using. In any case I am happy with *any* test docuemnt, and I don't mind testing python on the way ;-)
texexec --passon=-recorder test.tex
Yup, I have also seen Taco's email. That is good, the most important option in life ;-)
So I had an interfering version of math-eul.tex. When I deleted it, the test file compiled fine. Sorry for the noise.
Good to hear it resolved, and what is coming from you is never noise,
you *do* real testing, while others complain without testing and trying
to fix it themselves. Please continue with that noise!
Best wishes
Norbert
-------------------------------------------------------------------------------
Dr. Norbert Preining
participants (3)
-
Norbert Preining -
Sanjoy Mahajan -
Taco Hoekwater