| Age | Commit message (Collapse) | Author |
|---|
|
|
|
to be used later once it's in snaps.
|
|
eschew building the package is -n is mentionned.
Document that -S -n is heavily optimized for speed since it's mostly used
by dpb -R to figure out what to rebuild.
|
|
The actual bug reverted in 1.128 was from "make print-plist-libs"
which would invoke pkg_create -n -Q and filter out the libs: but
in that case, pkg_create would not be silent, thus yielding
reading plist|-/usr/local/lib/libpython3.9.so.0.0
to filter, which obviously wouldn't work.
So, turn on silent mode for -Q as well.
|
|
|
|
from Frederic Nowak, thanks !
|
|
show a "processing" message when we're actually grabbing data
(often from the net). This should be really quick in most cases,
but sometimes, it might take minutes (like for texlive), so that
we know what's going on.
|
|
|
|
|
|
and flag multiple subpackages as an error
|
|
an update.
|
|
okay sthen@
|
|
errors several times (for instance, if a file doesn't exist, its checksum
is not going to match, obviously)
|
|
don't do it again (would skew the file system usage a bit).
Also: we don't need to check for firmware if we're not actively deleting.
normal updates won't touch them.
|
|
the code that calls delete_handle already
|
|
several hundred copies of the same file (even thousands! I'm looking at you
sdcc), so matching through a list is O(n^2).
So hash the hashes on the actual file name so that we match directly if
the same name file didn't change, and use the first entry otherwise, since
the name won't match.
speeds up updating of texlive somewhat, and makes a huge difference for sdcc.
|
|
|
|
|
|
changes, and be more specific about @digital-signature and @signer, since
we only use the one scheme now.
|
|
(they probably won't mean a lot to most users but then they only
appear fleetingly apart from stuff like texlive)
|
|
* Say what happens when no pkg-name is given.
* Say "show the names and one-line comments" rather than just "information".
* Say which options cannot be combined with pkg-name arguments.
* Say which options are already the default in some circumstances.
OK espie@
|
|
OK espie@
|
|
any specific handling.
|
|
|
|
|
|
|
|
tweak @option always-update to allow parsing
@option always-update <hash_value_for_the_whole_plist>
once this is safely in snapshots, we can generate it
directly within PkgCreate.pm, compare it directly in
Signature.pm and get rid of the whole special case
of having to keep the whole plist around
|
|
(as it happens, some py-sip thingy got the treatment so I could test
that the code does something) !
|
|
|
|
|
|
|
|
trim content and add more explicit Xr to other documentation
|
|
- create a separate path when we don't want history
- add a dummy entry to see the threshold between changed and unchanged
in verbose mode
- document better the various stuff that gets separated
- error out in case we want history (by default) and anything goes wrong
|
|
we signed the packing-list separately. Now that we use signify to
create the signature without decompressing anything, it's completely
irrelevant.
|
|
It's just a progressmeter after all, there's no reason to error out instead
of simply devolving to "no progress meter".
(as noticed working with one of my students, Matthieu Fourre, whose
emulator or setup was somehow passing an "xterm-kitty" string)
|
|
|
|
|
|
|
|
whether we update or not, so we need to decorate the update_info with
the signer timestamp. It's the only place where an update_info actually
needs that stuff.
|
|
|
|
so I can show a percentage of cached names already handled.
|
|
optimize archive reading slightly: read one less header
if we skip things, and we don't need to check wanted for
emptiness if it didn't change.
|
|
that explain some of the more complicated stuff going on now
|
|
also fix a logic error (that's very unlikely to happen outside of broken
scenarios, but we still want to have things work correctly in that
context)
|
|
For update-info, we used to try to stop reading the CONTENTS file from
the stream as soon as we had enough, but
- most update-info will be cached
- the ones that won't are @option always-update, so they need the full
plist anyway
- most packages are signed, so we need a full 64KB of data to check the hash.
|
|
signature dates, just annotate the PackageLocation with the info,
and decorate the PackingList itself when we finished reading it from the
location
|
|
|
|
make the matching algorithm between the old and the new package more
precise: keep a list of matching checksums, instead of just one candidate.
For starters, empty files will always yield the same checksum, but it seems
some software (python for instance) loves installing multiple copies of
the same file. This does prevent 500 "name mismatches" out of 2500 in
python-3.9, for instance.
|
|
change the defines:
TEST_CACHING -> !NO_CACHING
TEST_CACHING_VERBOSE -> CACHING_VERBOSE
TEST_CACHING_RECHECK -> CACHING_RECHECK
okay with sthen@
|
|
|
