:7461740 Hi @egreg. Had to leave for a bit. Here is a M(n)WE that gives me the error. I'm currently looking into ltxkeys:

\documentclass[draft]{elsarticle}
\usepackage{ltxkeys}
\usepackage{standalone}
\begin{document}
\end{document}

Oh, it doesn't require elsarticle. Just article in draft mode does it too.

And switching the order between those packages fixes it.

But that doesn't work in my real document, because then I get a similar error related to pgfcore.

@JasperLoy Hello!

@JasperLoy Thank you and I wish you the same!

@egreg It's actually the catoptions package, as it turns out. New MnWE:

\documentclass[draft]{article}
\usepackage{catoptions}
\usepackage{standalone}
\begin{document}
\end{document}

catoptions does a lot of scary stuff.

@mhelvens Yes

@mhelvens See the defensive code I had to add to expl3

@JosephWright Do you have a specific reference?

@JosephWright Also, do you expect it could help me? Or is it very specific to defend expl3?

@JosephWright Is there any off-the-shelf solution to include vertical column seperator lines to Xsmallmatrix environments in mathtools?

@mhelvens Specifically to deal with a bad redefinition in catoptions

@percusse No idea: daleif is the mathtools expert

@mhelvens And many wrong ones, IMHO.

@JosephWright Oh OK. Thanks though.

@egreg Yes: does not keep the redefinitions properly compatible with the kernel

@egreg Wish I knew to stay away from ltxkeys before I used it as an integral cog to my own package.

@JosephWright do you have your moderator hat on? This new user has posted three quite similar questions within quick succession... tex.stackexchange.com/users/23809/piratepartypumpkin perhaps they're different, but I think they are very closely related

@mhelvens There are other keyval packages :-)

@cmhughes Taking a look

@JosephWright Yeah, but at first sight that one looked the best maintained with the nicest options. How could I know?

@mhelvens You did notice that pgfkeys is pretty well used? And that we are kind-of keen on l3keys working :-)

@mhelvens Heiko is also pretty good :-)

@JosephWright That's all relatively new information. When I started with my package I was still wet behind the ears.

You may remember some of it.

@egreg @JosephWright: Oh wise ones. Given the MnWE above, can you come up with some defensive measures?

@mhelvens Will have to look at what is broken in catoptions

@JosephWright I'm trying to do that right now. It's not easy.

@mhelvens Load catoptions after standalone

@cmhughes At least two of the questions are as you say linked

@JosephWright There's only one true original though.

@DavidCarlisle PStricks? :-)

@egreg That works in this particular minimal example. But with draft mode on, catoptions messes with several other packages too. I'd have to fix that too.

@mhelvens Yes, it certainly does

@mhelvens The problem is that it's definitions are defective, basically

@egreg Still, that's the approach I'll take, if there is no 'easy' way to fix this.

@mhelvens I never use Ahmed's packages. He tries to bee too smart (and the documentation is really difficult to read).

@egreg He's a clever guy, but he does as you say have a tendency to try to be too clever

@egreg I actually communicated with him about some bugs in ltxkeys. He seemed responsive.

@egreg I've had a lot of e-mail questions from him!

It looks like if I just manage to load catoptions last, things might be OK. I'll check.

@mhelvens That's likely, but beware \AtBeginDocument loading

(But it sucks, since I have quite a complex hierarchy of personal packages.)

@JosephWright In that case, I'll \AtBeginDocument catoptions. ;-)

@mhelvens The issue we had with catoptions came up in github.com/latex3/svn-mirror/commit/…

The problem was that it was looking ahead for tokens in a not-so-safe way

Yes, I see. If you try to hack TeX on such a fundamental level you can never be sure that you cover every case. I once redefined \newcommand, you may remember. (That actually seemed to work OK.)

@mhelvens I redefined \newcommand once as well:-)

@DavidCarlisle I needed to keep track of defined commands as well as redirect their definition to a global macro so I could play with first-class referencing environments. Still working on that, by the way.

@DavidCarlisle And I have some pages in my book where I explain how that redefinition works.

The problem there is that there are many other ways to define new commands which don't use \newcommand, so you still wouldn't cover every case.

@egreg Hmm probably it's easier to read the TeX than Italian:-)

@egreg @JosephWright By the way, loading my own package (which indirectly uses catoptions) last worked. Anyway, thanks for helping me find another pitfall to stay away from. I'm going to rewrite my concepts package to LaTeX3 at some point, so that'll fix that.

@DavidCarlisle It's from the bottom of page 170 to page 175 (almost full). I enjoyed some of your clever tricks when writing the explanation.

Yesterday I flagged an answer. Where can I see whether this is still in the pipes, or simply rejected?

@mafp I think you can see something in your profile about how many flags were helpful

@mafp this way:

tex.stackexchange.com/users/flag-summary/21591

@Joseph @JasperLoy Thanks. But it is odd: the flag was helpful, but the flagged answer is still there...

@mafp We very rarely mark flags as unhelpful

@JasperLoy yeah, the mod can agree with you but not take any action, e.g. because the question or answer was significantly improved

@egreg source code for TeX-in-SAIL should compile on PDP10 (as that was the development machine)

www-cs-faculty.stanford.edu/~uno/programs/texmf-sail.tgz

@DavidCarlisle Well, LuaTeX isn't there, nor XeTeX. :)

@egreg you don't need any of that funny foreign stuff

@DavidCarlisle Are you saying that I should emulate them in Plain TeX? ;-)

I just lost three days of work, blame my carelessness!

@kan there's only one thing that can be said: shit happens.

@tohecz So true. I am only hoping I can reconstruct everything.

Once again... facing this problem in emacs. Can someone help me please...

The problem is: I am using AuCTeX and C-c C-c used to compile default in pdflatex (I have explicilty said so in my .emacs file).

@kan Just punch in random C- and M- commands and Emacs will understand that you are lost :P

or it will write the article for you

But, now, for reasons I cannot comprehend, I am seeing default view instead of the instead of the standard thing I used to see....

@percusse better so!

@David might perhaps know! Please help me @David!

@kan I'm not sure I understand the question(and it's actually a while (years) since I really used auctex for compiling, I tend to just run latex on the command line:-)

The problem was (looks OK now): C-c C-c is not behaving well.

It gets to viewing instead of compiling.

@kan well that means it thinks the pdf is newer than the TeX

C-c C-c just means "do the next thing that needs doing"

@DavidCarlisle Oh, great!

@DavidCarlisle So true in this case.

@kan That will teach you: never argue with emacs

(I lost some work but pdf was still there... this time, someone played this game on my files)