Re: Bug#930219: ITP: node-dagre-layout - should other dagre users switch?
Changes from dagre to this fork are summarized at [0]. It looks like
Gitlab use this fork because Mermaid does [1] and the fork author is a
major Mermaid contributor [2].
A few packages currently use the original dagre(-d3), but none run its
full build process:
firefox/firefox-esr/thunderbird - embed already concatenated but
non-minified source devtools/client/shared/vendor/dagre-d3.js
snakemake - snakemake/gui.html loads dagre-d3 from upstream website
theano - starts from the full source, but doesn't use its build script:
runs browserify-lite + uglifyjs directly from d/rules
ssreflect used to embed dagre, but stopped doing so without apparent
replacement, possibly (I haven't checked) disabling the documentation
graph that uses it.
I previously [3] raised the question of whether dagre-* should be
packaged separately rather than embedded, without reply. I don't know
whether it would be practical or desirable for the above packages to
switch to this fork.
[0]
https://github.com/tylingsoft/dagre-layout#changes-compared-to-dagrejsdagre
[1]
https://gitlab.com/gitlab-org/gitlab-ce/commit/131e74d10dafbf2b781ab5d5517e42a18e20a587
[2]
https://github.com/knsv/mermaid/commit/7b935823da2058243dfc32f7c2a533ae233a9d1e#diff-8ee2343978836a779dc9f8d6b794c3b2
[3]
https://alioth-lists-archive.debian.net/pipermail/pkg-mozilla-maintainers/2017-September/029645.html
Reply to: