Feedback on Átila's Vision for D

Russel Winder russel at
Thu Oct 17 16:46:04 UTC 2019

On Thu, 2019-10-17 at 14:14 +0000, Atila Neves via Digitalmars-d wrote:
> As far as I can understand then, in your opinion cargo is better 
> than dub because:
> * It can use direct links to git repositories instead of 
> mandating the owners of those repositories explicitly register 
> them (that's all does, associate an email address 
> and package name with, more often than not, a git repo on github).
> * TOML vs SDL or JSON
> * The community rallies behind it
> * Details of where compilation artifacts are on the filesystem
> Is that an accurate summary?

It is a good summary of the points so far, certainly – assuming I have
understood the point about Dub and Git, Mercurial, and Breezy repositories.
The point about where compilation artefacts are is complicated and definitely
not black and white – minor changes to the way Dub handles things might make
it better than what Cargo does, especially for non-Dub tooling.

It is worth noting that it seems impossible to specify the application or
library version number in the dub.sdl file. It seems that Dub assumes
compilation in a Git repository, which would be a problem for source
distribution tarballs or zip files.

From what I can see both Cargo and Dub fail on updating generated source
files. An example: I generate a D module for Fontconfig library using DStep,
but it seems SCons is the only tool of Dub, Cargo, Meson, etc. that easily
copes with this build precursor. But I may be missing something.

It is worth noting that Dub and Cargo have more similarities than differences
overall, but this is probably why the differences can generate so much heat
and angst.

Dr Russel Winder      t: +44 20 7585 2200
41 Buckmaster Road    m: +44 7770 465 077
London SW11 1EN, UK   w:

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: This is a digitally signed message part
URL: <>

More information about the Digitalmars-d mailing list