[CommonHTML] equation label not properly aligned in large arrays with many labels

[pkra]

This is a continuation from mathjax/MathJax#1500, I'm afraid.

The minimal example I had posted seems to be fixed in both MathJax and mathjax-node (where I had run into this). However, a more complex example (the original one where I encountered this) seems fixed in the browser, but still broken in mathjax-node.

Below is the (pretty messy) example.

\begin{eqnarray}
  && \mathfrak{d}\mathfrak{d}' =\delta _{\mathfrak{d}, \mathfrak{d}'} \mathfrak{d}. \tag{57}\cssId{i}{}{}\\
  && \mathfrak{d}' E_n \mathfrak{d}=0, \quad {\mathit{if}}\ {\mathrm{sgn}}(\mathfrak{d}') =\pm , \tag{58}\cssId{ii}{}{}\\
 && \mathfrak{d}' F_n \mathfrak{d}=0, \quad {\mathit{if}}\ {\mathrm{sgn}}(\mathfrak{d}) = \pm , \\
 && \mathfrak{d}E_h [(\mathfrak{d}- E^{\theta }_{h, h} + E^{\theta }_{h+1,h+1}, {\mathrm{sgn}}(\mathfrak{d}) ) ] = E_h [(\mathfrak{d}- E^{\theta }_{h, h} + E^{\theta }_{h+1,h+1},{\mathrm{sgn}}(\mathfrak{d}))], \\
 && \mathfrak{d}F_h [(\mathfrak{d}+ E^{\theta }_{h, h} - E^{\theta }_{h+1,h+1}, {\mathrm{sgn}}(\mathfrak{d}))]\\
&&\qquad \qquad \qquad = F_h [(\mathfrak{d}+ E^{\theta }_{h, h} - E^{\theta }_{h+1,h+1}, {\mathrm{sgn}}(\mathfrak{d}))],\quad {\mathit{if}} \ h\neq n, \\
  && F_nE_n \mathfrak{d}- \mathfrak{d}' F_n E_n \mathfrak{d}= [\lambda _n]_v \mathfrak{d}, \quad \quad \quad \hspace{58pt}\! {\mathit{if}} \ \lambda =\lambda ', \epsilon = - \epsilon ' \neq 0, \tag{59}\cssId{iii}{}{}\\
 && \mathfrak{d}' F_n E_n \mathfrak{d}=0, \quad \mathfrak{d}F_n E_n \mathfrak{d}' =0, \quad \quad \hspace{56pt} \ {\mathit{if}} \ {\mathrm{sgn}}(\mathfrak{d}) =0, {\mathrm{sgn}}(\mathfrak{d}') =\pm , \\
 \ && (E_iF_j-F_jE_i) \mathfrak{d}=0, \quad \hspace{107pt} \ {\mathit{if}} \ i\neq j,\tag{60}\cssId{iv}{}{}\\
 && (E_iF_i-F_iE_i) \mathfrak{d}= [\lambda _{i+1}- \lambda _i]_v \mathfrak{d}, \quad \hspace{60pt} {\mathit{if}}\ i\neq n, \\
 \ && (E_iE_iE_j- [2]_v E_iE_jE_i+E_jE_iE_i) \mathfrak{d}=0,\quad \hspace{25pt}{\mathit{if}}\ |i-j|=1,\tag{61}\cssId{v}{}{}\\
 && (F_iF_iF_j- [2]_v F_iF_jF_i+F_jF_iF_i) \mathfrak{d}=0,\quad \quad \hspace{23pt}{\mathit{if}}\ |i-j|=1, \\
 \ && (E_iE_j-E_jE_i) \mathfrak{d}= 0, \quad (F_iF_j-F_jF_i)\mathfrak{d}=0, \quad \ \ {\mathit{if}}\ |i-j|&1,\tag{62}\cssId{vi}{}{}\\
 \ &&(E^2_nF_n+F_nE_n^2) \mathfrak{d}= [2]_v (E_nF_nE_n-E_n (v^{\lambda _{n+1} - \lambda _n+1} +v^{- \lambda _{n+1} +\lambda _n-1} )) \mathfrak{d},\tag{63}\cssId{vii}{}{}\\
 && (F_n^2E_n+E_nF_n^2) \mathfrak{d}= [2]_v (F_nE_nF_n- (v^{\lambda _{n+1}-\lambda _n-2} +v^{-\lambda _{n+1} + \lambda _n+2} ) F_n) \mathfrak{d}. 
\end{eqnarray}

[pkra]

Oh, I need to add two things.

a) I see the problem only when setting displayAlign and TagSide left.
b) the result then shifts a lot under browser zoom.

[dpvc]

Looks like the padding is getting lost in jsdom. Apparently the patch the I made for padding neglected to handle pixels without the px (e.g., node.style.padding = "1em 0"), which is legal, but deprecated.

I have patched my patch in the issue244 branch. I will need to submit a PR for CSSStyleDeclaration as well. Sigh.

[pkra]

Thanks, @dpvc!

[pkra]

Looks like this is now even more fixed upstream. Yay!

[pkra]

With MathJax 2.7-beta, the upstream fixes are now official and I think we can close this.