HOME kept across su without - made every detection probe raise PermissionError, and the run and the interface ended on a traceback. Detection now finds nothing there, says which folder this user cannot open and how to go on, on the first screen too.
A system_directory read from retroarch.cfg under a folder this user cannot open ended the run, and the interface before its first screen, on a traceback. It now gets the refusal --dest gets. On a handheld, an idle close on an error screen shows that error for a few seconds instead of waiting for a button that cannot come.
Path.exists() raised PermissionError from _missing_dest in line mode and from _choose in the interface. The run now names the folder and exits 1 (error event stage dest), and the interface asks for another folder.
install.py > log.txt at a terminal wrote the menu into the file and
waited for an answer nobody saw. A question now needs stdin and stdout on
a terminal, otherwise the run goes on as with --no-input and says so.
Path.exists() lets PermissionError through, so a BIOS subfolder without
search permission stopped --check and installs before any file. The file
is now reported unreadable and the run goes on with the others.
The test released the held downloads after a fixed pause. Under load the stop arrived after the release, all three files landed and nothing was stopped. It now waits for the run to say it is stopping.
An empty NO_COLOR turned colour off, which no-color.org reads as unset.
FORCE_COLOR and CLICOLOR_FORCE now paint a pipe. --check with several
platforms asks which to check rather than "Install for:", and
--list-platforms says when nothing is detected here.
A run stopped after 8 of 118 files said 118 were installed. The count
now comes from the downloads that finished, and a refused folder writes
one downloaded event per file, as the events table says.
_stable_hashes retried through the hash cache, so both retries returned
the first reading and the file was read once before the manifest run
stopped. The retries now read the file again.
On Windows the partial copy of a file of 16 MB or more is locked with msvcrt.locking, a LockFile region: the locking process cannot access that region through a second handle (Microsoft LockFile remarks). Hashing reopened the path, the read failed with a sharing violation, and every large file was reported as 'folder not writable' (issue 82). Both digests are now read in one pass through the handle that holds the lock. A read-only file already at the destination had the same misleading report: on Windows its attribute is now cleared before the verified copy replaces it, and a refusal that remains is named for what it is.
Opening and locking a partial download are two steps. Between them another run could finish, rename the same file into place and release its lock. The second run then locked the installed BIOS, read its full size as a finished partial and truncated it to download again. After the lock the partial's name must still point at the locked file, or the run downloads into a private copy.
Profiles, the wiki, comments, docstrings, command output and schema descriptions carried em-dashes, clause-joining semicolons, unicode arrows, box-drawing banners and check-mark glyphs. Each sentence was rewritten rather than having its character swapped. Hashes, paths, source_ref values and every cited file:line are unchanged, which a comparison of each parsed profile against the previous revision confirms.
validate_pr comments now say PASS and FAIL in words, and source_ref values in a profile's structured blocks render as code. test_prose_style fails if any of these comes back, in tracked files or in the generated site.
The cache of manifest hashes was keyed by the database's id without
holding the database, so a freed database replaced at the same
address got another scan's verified hashes back. The key now carries
the sha1 that database expects for the path.
Abdessamad Derraz