| git.druid.rocks | index | druid520 | mp | docs/ | reference/ | commands/ | fetch.btft |
docs/reference/commands/fetch.btft
[table class="topnav"]
[tr]
[td class="logotab"]mp[e]
[td][see name="index"]index[e][e]
[td][see name="guide/index"]guide[e][e]
[td][see name="reference/index"]reference[e][e]
[e]
[e]
[h1]mp fetch[e]
[h2]synopsis[e]
[code]
mp fetch <pkg[:slot]>
[e]
[h2]description[e]
pre-warms a port's fetch cache without building or installing anything -- runs just its [code]fetch:[e] phase, staged into its own [code]WD/fetch-cache/<canon>[e] directory, deliberately separate from the real build stagedir so it can never collide with a real install's own "already staged" check.
a port with no separate [code]fetch:[e] phase at all (every legacy [code]add.sh[e]/[code]del.sh[e] port -- its one script does fetch+build+install combined, nothing to pre-fetch in isolation) prints a note and exits cleanly; there's nothing to do, not an error.
the cache is per-canon (bare name plus effective slot), so two statically coexisting slots of one [code]pkg_name[e] (the [code]gcc-12[e]/[code]gcc-13[e] pattern) get separate caches rather than one falsely reporting the other as already fetched. fails if a cache for this canon already exists -- remove it by hand first (the message says exactly where).
[h2]examples[e]
[code]
mp fetch figlet
ok: figlet fetched into /usr/mp/fetch-cache/figlet.
[e]
[h2]see also[e]
[see name="reference/commands/install"]install[e], [see name="reference/commands/clean"]clean[e].