and installation (ditto update)
closes PR 14263
this may be a release branch candidate. I'm not sure myself, but the PR
sounds worrying.
If it should be considered, then this is the relevant diff (since more than
half of the diff is whitespace changes) to the file pkg/gui/main:
@@ -57,6 +57,8 @@
(define terminal #f)
(define (in-terminal-panel abort-label thunk)
+ (cond
+ [(or (not terminal) (send terminal can-close?))
(send dlg begin-container-sequence)
(when terminal (send terminal close))
(define t (in-terminal
@@ -71,7 +73,10 @@
(disallow-close)
(send dlg end-container-sequence)
(yield (send t can-close-evt))
- (allow-close))
+ (allow-close)]
+ [else
+ (message-box (string-constant install-pkg-dialog-title)
+ (string-constant install-pkg-not-rentrant))]))
Run tests in separate processes, support tests in
parallel, flag tests with non-zero exit codes or
stderr output as failing, add timeout support, etc.
Use the `--drdr` flag as a shorthand for DrDr-like flags.
The `--drdr` flag causes `raco test` to check for a `drdr`
submodule, then a `test` submodule, then run the module
directly. (The idea is that DrDr will eventualy try the
same sequence.) A test can declare an alternate timeout
through a `config` sub-submodule (and the idea is that
"props" will go away).
For example, make `raco exe --wrong` show "raco" instead of "racket"
as the error-reporting program. Normally, only `raco` and `setup-plt`
are implemented as "independent" launchers, because it's needed for
bootstrapping.
Now that the documentation-build phase of `raco setup` uses read-only
mode, contention between writing and reading sometimes (though
infrequently) triggers an SQLITE_IOERR_LOCK error. Change the SQLite
binding to expose that specific error, as well as SQLITE_IOERR_BLOCKED,
because a retry of the affected operation is appropriate.
Merge to release branch.
An invalid (needs-rollback?) transaction may correspond to a closed
underlying transaction. In that case, don't issue ROLLBACK command.
Using something like `raco pkg remove` in an incompletely
set up installation could lead to failures attempting to
access a database file that isn't there.
When `dynamic-require` is used to access an export that isn't a
variable defined in the same module, `dynamic-require` falls
back to `eval` in a fresh namespace, which can be expensive.
The new fast path handles the case that a variable is re-exported.
The new fast path is relevant to deserialization, which now
uses a submodule that re-exports from an enclosing module.
A change in `raise-reader-error` to use `srcloc->string` caused
tests to fail because `srcloc->string` doesn't show location
information if the source is #f.
