Fix version numbering

[chmreid]

This fixes inconsistent version number behavior as mentioned in #475 - also closes #475.

Before merging this pull request, the version number for the master branch should really be changed from "6.5.0" to something like "6.5.0dev" "6.6.0dev" (or similar) that will indicate it is different from the tagged version 6.5.0.

[amarjandu]

can we bump up the patch version of it? vs having 6.5.0dev

[amarjandu]

lgtm, this version file always gets messed up >.>

[chmreid]

can we bump up the patch version of it?

Do you mean 6.6.0dev or 5.6.1?

[DailyDreaming]

I'm fine with this.

Just for reference, the old way of distinguishing dev was a 0.0.0 version. I think at some point Jesse and Hannes decided to deviate from this (I'm not sure it's documented anywhere or that the old convention was either).

You have to go back a bit: https://github.com/HumanCellAtlas/dcp-cli/blob/fdb716e69491c2b797883966d1541ab338ca82af/hca/version.py

You should modify this line to catch versions ending in dev instead of catching 0.0.0:

dcp-cli/hca/cli.py

Line 57 in 22886df

if __version__ == '0.0.0':

That said, keeping the version as an rc and specifying it with dev sounds good to me.

[jessebrennan]

@DailyDreaming I've made no (*intentional) changes to how the versioning was done... I have done a couple releases, but I just followed the procedure in the DCP Ops Wiki. I find your vague and unwarranted suspicion is off-putting and befuddling.

[amarjandu]

@DailyDreaming I've made no changes to how the versioning was done... I have done a couple releases, but I just followed the procedure in the DCP Ops Wiki. I find your vague and unwarranted suspicion is off-putting and befuddling.

If this fix means that we don't have to keep changing this file back to 0.0.0 that would be perfect. I've committed this file before to a random version number, it always manages to sneak into the commit.

[DailyDreaming]

@jessebrennan I'm sure it was accidental (and something I've done before too). No suspicion or otherwise intended! As Amar mentioned, the version in master was kept at 0.0.0 but it gets rewritten and accidentally included in commit changes, so we used to have to be wary and make certain that it was changed back. I've changed it accidentally too, and because there's no documentation around it, I'm sure you (or someone else) accidentally changed it without realizing it.

I think setting it to the actual version with an rc or dev modifier makes a lot more sense.

[codecov-io]

Codecov Report

Merging #476 into master will increase coverage by 0.7%.
The diff coverage is 50%.

@@           Coverage Diff            @@
##           master    #476     +/-   ##
========================================
+ Coverage   85.19%   85.9%   +0.7%     
========================================
  Files          41      41             
  Lines        1844    1986    +142     
========================================
+ Hits         1571    1706    +135     
- Misses        273     280      +7

Impacted Files	Coverage Δ
hca/cli.py	63.39% <0%> (ø)	⬆️
hca/version.py	100% <100%> (ø)	⬆️
hca/util/__init__.py	89.23% <0%> (+2.24%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update ca60108...811ca93. Read the comment docs.

[chmreid]

Set version to 6.6.0dev and updated the version check function to see if version ends with dev