January 2023 NEWS Update.#1557
Conversation
whitwham
force-pushed
the
news_jan_2023
branch
from
dd8e13c to
9879637
Compare
NEWS
Outdated
| * Fix a bug in the codec learning algorithm for TOKA. The name tokeniser has a | ||
| rANS vs Arithmetic coder choice as a parameter (in the "strat" variable). We | ||
| lacked this distinction when learning which method works best, so in the | ||
| choice of toka (tok3+arith) vs bzip2 vs gzip etc we selected tok3 and switched | ||
| back to strat 0, disabling the arithmetic coder.his only affects archive mode, | ||
| or where requested eg "samtools view -O cram,version=3.1,use_arith". | ||
| (PR#1559) |
There was a problem hiding this comment.
I think less is more with NEWS. If we're too detailed people simply won't bother to read it anyway, so it should be as brief as possible. The PR numbers let people jump to the full details if it's something they're interested in.
For intricacies like this I'd just paraphrase it in a single sentence. I'd suggest
"Fixed a bug in the codec parameter learning for CRAM 3.1 name tokeniser (PR#1559)"
NEWS
Outdated
| * Catch errors from bgzf_getline in hts_readlist and hts_readlines. There were | ||
| a couple of places where the returned error was ignored. |
There was a problem hiding this comment.
Similarly to below, we can cull the second sentence IMO.
NEWS
Outdated
| * Fix error code returned by bcftools. Error code was returning 0 (success) | ||
| even on a fail. Replaced by -1. | ||
| (PR#1504, thanks to Lilian Janin) |
There was a problem hiding this comment.
This is both too wordy and too content free. I understand where it comes from as it's the commit messages, but for PR#1504 all the actual explanation is in the conversation and not the poorly worded commit messages.
Maybe replace with:
"Bcftools no longer returns 0 exit status when encountering bgzf decode errors".
It may also be worth just bundling up a whole bunch of related things together. Eg
"Improved bcftools detection and reporting of bgzf decode errors"
That covers #1504, #1529, and #1554. Our readers don't really need to see the finer details of how the bugs were reported and fixed separately, given they're essentially all the same route cause of missing bgzf error handling.
NEWS
Outdated
| or where requested eg "samtools view -O cram,version=3.1,use_arith". | ||
| (PR#1559) | ||
|
|
||
| * CRAM improvements. Fixed crash with multi-threaded CRAM.Fixed a bug in the |
There was a problem hiding this comment.
"multi-threaded CRAM.Fixed a bug" - missing space after full-stop. (Or two spaces, but that'll be my ancient brain.)
|
Looks gtg, apart from an additional space. :) |
2 participants
Putting up the first draft now. Will put the final one in just before the release.