Releases · immortalvm/dbspec

25 Apr 09:50

v1.3.1

Release Notes for v1.3.1

Improves how the source is displayed/logged in error messages.
Allows embedded PowerShell scripts under Windows (by adding .ps1 suffix
to temporary script files).

We have also made some small changes to the HBR example.

Assets 3

26 Mar 14:34

v1.3.0 Latest

Latest

We now strip (at most) one trailing newline from the result (i.e.
standard output) of non-SQL external scripts, saving many uses of
.stripped.
In a loop, when the row has fewer columns than the variables we bind,
the extra variables are made undefined (rather than keeping their
previous value).

Handle non-ASCII characters and CR correctly.
Print line numbers error messages starting from 1 (not 0).
Wait for SIARD process to stop to avoid exceptions for small databases.
We now prevent nested iteration over the same result set (which has
never worked as expected).
Fix parser issue leading to wrong SIARD metadata in some cases.

Better warnings when generating SIARD files.
Signal failure when metadata is specified for database objects not
included in the generated SIARD file.
Show warning when metadata is missing for schemas, tables, views and
table columns in the generated SIARD file
Prefix every log message with timestamp [hh:MM].
Used system specific line separator in log messages.
Show current variable bindings when assertions fail.
Improved failure messages.

.as_integer now accepts initial and trailing whitespace characters,
saving many uses of .stripped.
The values of 1x1 row sets can now be directly accessed using
.stripped and .as_integer.
Strings can now be compared to row sets and iterated over like row sets,
in which case it is assumed to consist of tab separated string values.
(However, .size can still only be used for actual row sets.)
Strings and row sets can be compared with == and !=. (NB. If r==a
and r==b for some row set r and strings a and b, we may still
have a!=b due to different line endings.)
The development option -use-existing-SIARD has been added for when
adjusting the metadata fails after the SIARD file has been generated
(which can be very time consuming).

The existing examples have been adjusted and updated.
A new example has been added: hbr/befolkning.dbspec.
(Alas, it uses a data dump which is not generally available.)

Assets 3