This status describes a state where the head branch is mergeable and
technically not blocked per base branch requirements, but it does have
non-passing checks.
Conditions prohibiting a regular merge: BLOCKED, BEHIND, DIRTY.
Conditions triggering a regular merge even if `--auto` was set: CLEAN,
HAS_HOOKS.
Note that UNKNOWN status does not trigger either of the conditions.
This removes the false equivalence between GH_CONFIG_DIR and
XDG_CONFIG_HOME. These settings do not have the same effect and should
not be used for the same purposes.
Also remove the documentation about what `XDG_*` settings do. We simply
conform to the XDG Base Directory Specification, but will not document
it. It's likely that users of these environment variables already know
what they do.
If GH_CONFIG_DIR is set, don't auto migrate the config file. This fixes
the situation where the path given via GH_CONFIG_DIR does not exist and
the cli attempts to migrate an existing config to that location.
Fixes#3837
The first paragraph uses single quotes when referring to shell arguments
and variables, but the rest of the docs use double quotes. This commit
switches to using single quotes throughout the docs.
I prefer to use single quotes inside string literals because Go uses
double quotes to define a string literal.
On non-Unix-like shells like Windows Command Prompt, single quotes are
handled differently. You need to define aliases using double quotes
instead of single quotes.
I added an inline example to illustrate the quotes. The example is
formatted as inline code blocks in Markdown. Unfortunately, because Go
uses backticks for raw string literals, I needed to do some rather ugly
string concatenation in order to get the backticks included in the doc
string.
This also rearranges the notes so that the platform specific notes are
at the end of the documentation.
This enforces strict separation between serialization structs used for
repository creation payload with respect to whether GraphQL or REST was
used. Before, a field added to a GraphQL payload would leak to REST
payload (and vice versa).
we don't ever check the return of Fprintf anywhere else in the codebase
so doing it here suggests that it's a special case. if it's something we
should be doing we can circle back and do it more consistently.
Upgrades gojq, the library that powers the `--jq` filter flag for JSON.
This upgrade is purely housekeeping and not to address any particular
issue reported with gh.
https://github.com/itchyny/gojq/releases/tag/v0.12.4
