optidef messes with overbrace somehow

[roversch]

The following LaTex code

\documentclass{article}

\usepackage{amsmath}
\usepackage{mathabx}
\usepackage{optidef}

\begin{document}

\begin{equation*}
\overbrace{T + H + I + S \times U + N + D + E + R}^{\mathrm{top}}
\end{equation*}

\end{document}

results in

If I switch the lines \usepackage{mathabx} and \usepackage{optidef}, it returns the correct overbrace rendering:

Note that if I don't include amsmath, it does not even compile. ('command \iint already defined' etc). My guess is that somehow, optidef relies on amsmath and is unhappy with mathabx being included first.

[jeslago]

Thanks for posting Robin! I will take a look at the clash and see how can be avoided.

[bonanza123]

I guess the problem is not in optidef but rather the fact that mathabx overwrites the definition of the braces etc. A reasonable way would be to load only those glyphs form mathabx that you need, via \usepackage[ putYourClassHere ]{mathabx}, cf. the mathabx documentation.

[jeslago]

Thanks for that answer. I also think that might be the issue, but I would take a look to the optidef code if something can be done to improve compatibility with mathabx

[roversch]

Thanks for the help! If mathabx is the problem, why does it work if you switch the statement order? I would expect it to give wrong results in both cases.

[roversch]

@bonanza123 thanks for that hint by the way, but \usepackage[mathx]{mathabx} also overwrites \neq, making it unusable.

[bonanza123]

There are a lot of packages where order matters (hyperref and cleveref for example). Lets ask the other way around, why do you need mathabx to begin with?

[roversch]

Okay, good point. I need mathabx for 'widecheck'.

On Fri, 13 Jul 2018 at 14:42 bonanza123 notifications@github.com wrote:

There are a lot of packages where order matters (hyperref and cleveref for example). Lets ask the other way around, why do you need mathabx to begin with?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/jeslago/optidef/issues/17#issuecomment-404822231, or mute the thread https://github.com/notifications/unsubscribe-auth/ABXrjLTvPXig8elOBjaO-TRUkzIpDY2iks5uGJWZgaJpZM4VOpoX .

[bonanza123]

Ahh, ok: I guess the problem with optidef is that it loads mathtools (which loads amsmath IMHO).

But for you, check out this: https://tex.stackexchange.com/questions/44235/is-there-a-way-to-do-an-upside-down-widehat/44251 regarding the alternatives to mathabx for that glyph.

[roversch]

That makes sense. Yeah, I saw that SO post before. Thanks for the clarification!

[jeslago]

I am closing this issue. I will add a section in the documentation to indicate the possible issue when using optidef with mathabx and how to solve it. From my side, there is nothing I can do in the optidef package.