Consistency and FAQs

[armintoepfer]

Consistency

• Use - for no available packages. Easier on the eyes.
• bioconda lowercase if not at start of sentence.
• *-Packages

Clarify conflicts of pb-dazzler

FAQ: No PacBio tech support. Use issue tracker

FAQ: Clarify which tool version to use, bioconda vs SMRT Analysis.

[pb-cdunn]

I guess that's fine. But I'd like to track which packages already work with py3. Hmmm...

... I think we need an Issue. There, we can use check-boxes to track our work on py3-compatibility.

[armintoepfer]

Will add a link to your issue

[pb-cdunn]

pypeflow and falcon can also conflict.

Actually, someone updated pypeflow over the weekend, and it happens to match the code in pb-falcon. One idea: turn those into trivial, immediately crashing code, so nobody will use them. Then, deleted their recipes (which I've seen, so I think it's possible).

Unfortunately, we cannot do that for DALGINER/etc because we do not actually own those. So we need a check in Falcon to test early whether we have our own daligner suite..

[pb-cdunn]

Btw, I learned something annoying when I looked into the pypeflow package. These are equivalent ways to obtain source trees from github:

1. https://github.com/PacificBiosciences/pypeFLOW/archive/{{ version }}.tar.gz
2. https://github.com/PacificBiosciences/pypeFLOW/archive/{{ commit }}.tar.gz

Unfortunately, commit/version goes into the directory name within the tarball. (And in fact, the directory name does not match the tarball filename; rather, it is prefixed with the repo name. But that's not the problem.) So these 2 100% identical source distributions have different sha256 checksums. That makes equivalence hard to prove. But I did verify it myself.

[pb-cdunn]

(Does my comment in python versions appear in this review. Let's see...)

[mpkocher]

Your usage of "SMRT Analysis" is confusing to me. "SMRT Analysis" is a subcomponent of SMRT Link. Similarly, "smrt-tools" is a subcomponent of "SMRT Link" that is used by several other subcomponents (e.g., SMRT Analysis, RunQC).

"SMRT Analysis build" is an RS-era construct. There is no (or there shouldn't be) a formal SMRT Analysis (only) build. There are only "SMRT Link" and "SMRT Link tools" builds.

[armintoepfer]

Thank you. Will push a fix

[pb-cdunn]

Remember to click "Delete branch" after merging, if desired.

[armintoepfer]

Yeah, in bitbucket I have the checkbox activated per default :) Thanks for the reminder