Yurii Rashkovskii
Yurii Rashkovskii
Even though indexing primitives are promised in our documentation, we haven't published them yet, leaving this to users. Proposed solution: before addressing particular index types (equality, ordered, etc.), design a...
``` PumpkinDB> [[DUP] TRY DROP 1 ASSOC?] READ. Error: "No transaction" "" 0x08 ``` In this situation, `DUP` without a stack generates an error which is caught by `TRY` but...
What would you do if you want to allow writing data with certain properties only once? Proposed solution: provide in-memory advisory locking primitives.
Parsing and analyzing PumpkinScript in runtime comes at a cost and can't be nicely optimized by the compiler. Proposed solution: develop a CPS (Continuation Passing Style) interpreter. The scope of...
This issue is mentioned in ad6da5243d4f5804592d9bd7fa67306e83a53f63 There was a realization that `Env
Previously, doctests were invoked against different feature sets. However, after #203, because the project was split into multiple crates, it is less obvious how to test all feature sets. cc...
In certain cases, especially when builtins are used, one module might depend on the presence of another module. Lets say, mod_storage's builtins can make use of `NIP` or `STARTSWITH?`. However,...
Currently we prohibit nested writes (#230) because they don't make sense until we support lmdb's child transactions Proposed solution: add support for lmdb's child transactions.
Currently, the server will not send the topic name over the TCP connection. That is done in order to avoid imposing rigidity and overhead, so that clients can elect how...
There are certain cases, like in-memory indexing or testing, where NOT storing data to disc between restarts would have been useful. (Also, MUMPS has this distinction as well in a...