Increase maximum diff size for notebook files

Everyone can contribute. Help move this issue forward while earning points, leveling up and collecting rewards.

• Close this issue

Apologies if this is a duplicate - I found this CE issue that was closed six months ago, and after reading through felt that it was more appropriate as a new issue/proposal.

Description

We're running 9.0.0-ee. One of our use cases is that we use GitLab to review work in Jupyter notebooks. When there are plots in the notebooks, the notebook format includes the plots as base64 encoded png data eg (truncated, with line breaks added to make it easier to see)

"data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAArgAAAHxCAYAAACGf88WAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAPYQAAD2EBqD+naQAAIABJREFUeJzt3X
t0VOWh/vEnXGYmKFQMSbgUysVWkoBJyMWiBlpUrELg\nnAY81loNoqkaokdra7BeEKtosSDlVogUEWtLmrRVpAuPVisSaIEESJbBI8EEEyRhYgWpmckE2
L8/\nPMzPIQSyhyQT3nw/a7Fced+9Z7/7cU982O5MwizLsgQAAAAYoluoFwAAAAC0JQouAAAAjELBBQAA\ngFEouAAAADAKBRcAAABGoeACAADAKBR
cAAAAGIWCCwAAAKNQcAEAAGAUWwW3trZWd911l5KSknT1\n1VdrzZo1/rmamhrNmDFDiYmJmjx5soqKigL23bJli9LT05WQkKDMzExVV1e3zRkAAAAA
X2Gr4N53\n33264IIL9Oc//1kPP/ywnn/+eb311luSpHvuuUdRUVEqLCzUlClTNGvWLNXW1kqSDh48qOzsbGVk\nZKiwsFB9+/ZVdnZ2258NAAAAurwwy7
s1mz4+eefKzU1Va+//rouueQSSdK9996rqKgoXXPNNbrn\nnnu0detWOZ1OSdKMGTOUlJSkWbNmadGiRSouLtZLL70kSfJ6vbryyiv1m9/8RikpKe10agAA
AOiK\nWn0H1+VyKTw8XIWFhTp27Jg++ugjlZSUKCYmRrt371ZcXJy/3EpSUlKSdu3aJUkqLS0NKLIul0ux\nsbHauXNnG54KAAAAYKPgOhwOPfbYY/rDH/
6g+Ph43XDDDRo3bpwyMjLkdrsVFRUVsH1ERITq6uok\nSYcOHWo2369fP/88AAAA0FZ62Nl43759mjBhgmbOnKkPP/xQTz75pMaOHSuPxyOHwxGwrc
PhkM/n\nk/TlIwlnmm+N5ORkNTY2NivKAAAA6BzcbrccDod27NgR0nW0uuBu3bpVBQUF2rRpkxwOh2JjY1Vb\nW6vly5dr7NixOnz4cMD2Pp9PLpdLkuR
0OpuVWZ/Ppz59+rR6oT6fT8eOHVNjY2Or9wEAAEDHaWpq\nUit/

The result of this is that the files end up big, and almost always won't be displayed in a merge request, making it impossible to add inline comments.

Obviously the max size can't be unlimited, but in moving to gitlab from github, it's apparent that their threshold is higher than yours, so whereas we would rarely run into this issue there, we almost always see it here, and end up having to strip the output from notebooks so we can review them in gitlab. This isn't a big deal until someone forgets this extra step, and then the MR is held up while reviewer has to wait for the author to do this before the review can proceed, or alternately leaves non-inline comments which make the iterative review process that much harder.

Based off the interest in https://gitlab.com/gitlab-org/gitlab-ce/issues/2508 I think our use case may not be unique.

Proposal

I'm not sure what the exact solution should be from a technical perspective - whether the technical limit could be increased, or whether there should be some special handling of notebook files, but being able to view a diff and add line by line comments to these larger notebook files would be the end goal.