A problem in submitting .pdf files to the arXiv.org

[kderakhshani]

Dear Friends,

As you know, one of the best preprint servers for physics papers is arXiv.org, and almost everybody desires to post her/his paper on this server so that it can be commented/corrected/improved by peer reviewers before publishing.

However, the price is that arXiv.org is somewhat strict. For example, it compels the submitter to post all of the latex package of the paper (including figures, macros, styles, etc.) if the pdf file is generated by latex. The rationale is that, this way, arXive can process it by its own latex compiler and generate the pdf version along with other formats (ps, ...). If you want to upload your paper in pdf format that is generated by your latex compiler (without your latex files) it will be rejected. And if you want to post your latex files that is not compatible with their latex versions/styles you will receive a pool of errors.

My problem is that I have prepared a version of my paper (accepted by The Astrophysical Journal) using the document style emulateapj.cls. This document class allows you to process your paper very similar to the published format and is very decent and good looking. But arXiv.org could not process it and delivered to me a hell of a lot of errors.

Is there any way to fix or bypass this problem?

Thank you very much for your help.

 

[DrClaude]

Why don't you just upload emulateapj.cls with the rest of your files?

 

[Ben Niehoff]

Yeah, I'm pretty sure you can just upload the .cls file along with the rest of your files.

arXiv appears to be using a standard TexLive installation, and I think they keep it pretty up-to-date. So it should have all the standard packages. Using a custom class file is not an "incompatibility"; it just means that you need to supply the class file. Same with any custom style files you may be using.

 

[AlephZero]

arXiv appears to be using a standard TexLive installation, and I think they keep it pretty up-to-date. So it should have all the standard packages.

Hmm... emulateapj is in my TexLive installation. (It's TeXLive 2013, but I don't bother to keep it right up to date).

Maybe arXiv don't allow arbitrary document classes for some reason (e.g. to try to maintain an approximation to a uniform format for all documents)?

 

[Ben Niehoff]

According to this page, they are on TexLive 2011. So not exactly up-to-date:

http://arxiv.org/help/faq/mistakes

 

[kderakhshani]

Dear Friends,

Thank you all for your replies.

However, as I (maybe somewhat implicitly) said, I did upload all files, including the .cls.

When I asked help from the arXiv admins by email, they answered as the following:

"The relevant error is:

Extra alignment tab has been changed to \cr.
<template> ...ip \tabrightsep \relax \endtemplate

l.197 ...&$26$&$1.9$&$20.43$&$6.24$&$16.47$&$126$&

?
! Emergency stop.
<template> ...ip \tabrightsep \relax \endtemplate

l.197 ...&$26$&$1.9$&$20.43$&$6.24$&$16.47$&$126$&

Output written on Apj87458R4.dvi (4 pages, 27500 bytes).
Transcript written on Apj87458R4.log.

This is a real error that you'll have to correct. Carefully count the number of
columns that you've created versus the number expected."I'm not sure that I understand it.

Would you please help me more?

My paper has been published online on Feb 12 and on paper on March 1, but I couldn't submit to arXiv.org as yet !

Thanks a lot.

 

[DrClaude]

Could you please post the source file lines around line 197?

 

[Ben Niehoff]

Are you sure this compiles without error on your own machine? Because it looks like you've set up a tabular environment incorrectly.

Without your source files in front of me, it's hard to figure out what is wrong (and I don't have time to look at it anyway). However, from the error lines you quoted, it looks like there is an extra & at the end of the line. Hence why the error complains about it being changed to \cr (carriage return).*

* LaTeX errors are always terribly worded. It takes a bit of decipherment to figure out what you actually need to fix.

 

[AlephZero]

Are you sure this compiles without error on your own machine?

If the OP uses a LaTeX editor that hides the messages, he/she might never have noticed the problem. If LaTeX recovered from the error, the output might be what was expected.

The simplest cause of the problem is that the & and \\ characters in the .tex file are inconsistent with number of columns in the table.

According to this page, they are on TexLive 2011. So not exactly up-to-date:

And worse, there is a bug in the TexLive updating system which means it can't update itself from 2011 to 2013 - it tries, but crashes. I discovered that last year after I found 2013 had some new toys that I wanted to play with. The fix was a complete new install of 2013, which would probably have some consequences for arXiv's software because TexLive have changed the system's directory structure.

 

[Ben Niehoff]

And worse, there is a bug in the TexLive updating system which means it can't update itself from 2011 to 2013 - it tries, but crashes. I discovered that last year after I found 2013 had some new toys that I wanted to play with. The fix was a complete new install of 2013, which would probably have some consequences for arXiv's software because TexLive have changed the system's directory structure.

I'm running TexLive 2013 and I have not had any problems with arXiv. The directory structure should not cause files to fail to compile, since standard packages are just referred to by package name and not the full path.

But yes, if arXiv wants to upgrade to TexLive 2013, they'll have to reinstall from scratch. I don't think this should be a major problem, though...it takes what, maybe an hour or two?

 

[AlephZero]

I'm running TexLive 2013 and I have not had any problems with arXiv.

So I guess you aren't using any packages in TL2013 which were not in TL2011 (unlike the OP).

The directory structure should not cause files to fail to compile, since standard packages are just referred to by package name and not the full path.

Agreed the new structure shouldn't cause any problems with TexLive itself (assuming TL hasn't created any bugs by changing it). But if arXiv have their own software to automatically generate different formats, that may cause some issues. IIRC there have been some changes to the "core" executables like PDFLaTeX etc.

But yes, if arXiv wants to upgrade to TexLive 2013, they'll have to reinstall from scratch. I don't think this should be a major problem, though...it takes what, maybe an hour or two?

That depends how much regression testing they want (or need) to do. Trying to stay completely up to date is sometimes like pushing a piece of string. There are updates to some component or other in TexLive pretty much every day. The strategy of "if the version you have does what you want, being out of date doesn't matter" sometimes works best!

 

[kderakhshani]

Dear DrClaude,

Sure. The following are the lines 178-206: (line 191 is the ‎\begin{tabular}{cccccccc}‎ )

 ‎\begin{figure}‎
‎\includegraphics[scale=0.7]{images/LOSVD_ov_del_ratio.eps}‎
‎\caption{Ratio of difference overall line-of sight velocity dispersions of the simulated model globular cluster with a mass of $M=7.5\times10^5 M_{\odot}$ and a stellar mass-to-light ratio $M_*/L=2.0$‎, ‎as a function of the galactic external field and for various internal fields‎. ‎$\sigma_{los,N}$‎, ‎$\sigma_{los,M}$‎, ‎and $\sigma_{los}$ are the Newtonian‎, ‎isolated MOND‎, ‎and MOND with an external field overall line-of-sight velocity dispersions‎, ‎respectively.}‎
‎\label{fig_5}‎
‎\end{figure}‎‎\begin{table*}[htbp]‎
‎\caption{The best-fit simulated models for NGC 2419 and their goodnesses of fit.}‎
‎\label{tab2}‎

‎\begin{center}‎
‎\begin{tabular}{cccccccc}‎
‎\hline‎
‎Model&$M$&$r_h$&$M_*/L$&$\mu_V(0)$&$\sigma_{los}(0)$&$\rho(0)$&$\chi_{total}^2$\\‎
‎&$(10^5M_{\odot})$&$(pc)$&&$(mag/arcsec^2)$&$(km/s)$&$(M_{\odot}/pc^3)$&\\‎
‎\hline‎
‎Newtonian Plummer&$9.0$&$25$&$1.9$&$20.24$&$6.53$&$26.68$&$48$\\‎
‎MOND+EFE&$8.0$&$25$&$1.8$&$20.27$&$6.25$&$25.10$&$119$\\‎
‎MOND-isolated&$8.0$&$26$&$1.9$&$20.43$&$6.24$&$16.47$&$126$&‎
‎\hline‎

‎\end{tabular}‎
‎\end{center}‎

‎\tablecomment{{\bf Notes}‎: ‎Newtonian Plummer‎: ‎The best-fit Plummer sphere in the Newtonian equilibrium; MOND+EFE‎: ‎The best-fit model in the MONDian equilibrium including a Galactic external field $0.14a_0$; MOND-isolated‎: ‎The best-fit model in the MONDian equilibrium in isolation‎. ‎The fit parameters are the mass $M$‎, ‎the half-light radius $r_h$‎, ‎and the stellar mass-to-light ratio $M_*/L$‎. ‎The derived parameters are central surface brightness $\mu_V(0)$‎, ‎central line-of-sight velocity dispersion $\sigma_{los}(0)$‎, ‎and the central mass density $\rho(0)$‎. ‎The $\chi_{total}^2$ is the minimized sum of $\chi_{\mu_V}^2$ and $\chi_{\sigma_{los}}^2$.}‎

‎\end{table*}‎

Thanks.

 

[DrClaude]

Just remove the extra & at the end of

MOND-isolated&$8.0$&$26$&$1.9$&$20.43$&$6.24$&$16.47$&$126$&‎

(the table has 8 columns).

 

[Ben Niehoff]

Also, you know, you can leave spaces around the &'s to make it easier to read, i.e.

\begin{tabular}{cccc}
1 & 2 & 3 & 4 \\
5 & 6 & 7 & 8
\end{tabular}

Then you'll be less likely to miss something like that. The spaces won't be printed.

 

[kderakhshani]

Dear Repliers,

I made the correction you suggested. Apparently that error is fixed, but I've received the attached response from the arXiv processor. My tex file is also attached


Would you please help me again?

Thanks a lot.