search

PeteHaitch / DelayedMatrixStats

A port of the matrixStats API to work with DelayedMatrix objects from the DelayedArray package
Other
15 stars 7 forks source link

Please use consistent signing in version numbers #88

Closed tillea closed 2 years ago

tillea commented 2 years ago

Hi, when I updated the Debian packaged version of DelayedMatrixStats I realised that the versioned dependency from the Matrix package was specified as "(>= 1.5.0)" while the Matrix package you are referring to is actually 1.5-0 (with a '-' sign, not a dot). I'm perfectly aware that R is sorting '.' and '-' is sorting equally. However, in other environments where this is not the case it is asking for trouble when using divergent spelling of the version string. That's why I applied a patch to the Debian package which boils down to

-        (>= 1.5.0), S4Vectors (>= 0.17.5), IRanges (>= 2.25.10)
+        (>= 1.5-0), S4Vectors (>= 0.17.5), IRanges (>= 2.25.10)

inside the DESCRIPTION file. It would be great if you could apply this to the next release. Kind regards, Andreas.

PeteHaitch commented 2 years ago

Thanks for bringing this to my attention, @tillea. I didn't know there were R packages that used '-' as a delimiter in version numbers. I've updated it in the development and release branches of Bioconductor. New versions should be available via the Bioconductor build system in the next 24-48 hours.

tillea commented 2 years ago

Am Tue, Oct 11, 2022 at 02:33:42PM -0700 schrieb Peter Hickey:

Thanks for bringing this to my attention, @tillea.

Thanks for closing this quickly.

I didn't know there were R packages that used '-' as a delimiter in version numbers.

Several years ago we replaced '-' by '.' but this turned out as a failure for several reasons.

I've updated it in the development and release branches of Bioconductor. New versions should be available via the Bioconductor build system in the next 24-48 hours.

There is no hurry needed actually since I've solved the issue with the patch. For future releases your change is very convenient.

Thanks a lot for the cooperation, Andreas.