StataCorp Stata 17 (Revision 08 Mar 2023)

1. collect and etable have improved support for estimation results from
menl and mixed. Variance components in the reported metric are now
put in the collection. The old behavior of collecting the variance
components in the estimation metric is now available with option
estmetric.

2. frame drop previously accepted only one frame name as argument. It
now accepts a list of frame names and supports Stata wildcards * and ?
in names: * matches zero or more characters, and ? matches exactly
one character.

3. The Java Development Kit that is redistributed with Stata is now
updated to version 17.0.6+10-LTS acquired from Azul Systems.

4. cmmixlogit and cmxtmixlogit, when used with option intmethod(random)
and when more than 20 variables were specified in option random()
(where the count includes factor-variable levels), erroneously exited
with an uninformative error message. This has been fixed.

5. collect export to Microsoft Word or PDF, when a cell border or margin
width was specified without units, used inches when it should have
used points. This has been fixed.

6. collect layout, when specified with a dimension that shows the title
but hides all the levels by collect style header, exited with an
unhelpful error message. This has been fixed.

7. The 10jan2023 update introduced a bug in import sas when reading
Unicode data. In this case, import sas exited with error message
"error reading file", r(692). This has been fixed.

8. jdbc load exited with error message "failed to load resultset column",
r(683), when loading a string column from an SQLite database. For a
TEXT/VARCHAR column, the SQLite JDBC driver incorrectly returned a
column size of 0, which triggered the Stata error. jdbc load now
stores TEXT/VARCHAR columns with 0 column size as strLs.

9. putdocx table, when inserting an estimation table using putdocx table
tablename = etable[(#_1 #_2 … #_n)] into a docx file and when
title() or note() option is specified to add a title or a note to the
table, would exit with uninformative error message "argument out of
range" if the length of the table name tablename exceeded 11. Now the
length of tablename is extended to 24, and a more informative error
message, "invalid table name; table name too long", is issued if the
length exceeds the limit.

10. putpdf table, when inserting an estimation table using putpdf table
tablename = etable[(#_1 #_2 … #_n)] into a pdf file and when title()
or note() option is specified to add a title or a note to the table,
would exit with uninformative error message "argument out of range" if
the length of the table name tablename exceeded 12. Now the length of
tablename is extended to 24, and a more informative error message,
"invalid table name; table name too long", is issued if the length
exceeds the limit.

11. svyset dialog box failed to populate its controls for the fay(),
bsn(), and calibration properties from the current survey settings.
This has been fixed.

12. In the Data Editor, using the Paste Special dialog could cause Stata
to crash. This has been fixed.

13. (Mac) In the Do-file Editor, highlighting all matches is now faster,
but highlighting all matches for very large files can still be very
slow. For performance reasons, the Do-file Editor will not highlight
all matches if the document size is greater than 1 MB.

14. (Mac) In the Do-file Editor's Find bar, setting "Highlight all
matches" was ignored, and all find matches were always highlighted.
This has been fixed.

15. (Mac) Some Unicode extended Latin characters such as a dotless "i"
could not be entered in the Command window or the Do-file Editor.
This has been fixed.

16. (Mac) macOS Ventura would not display Quicklook previews of Stata text
files such as do- or ado-files. This has been fixed.

17. (Mac) The Viewer scheme colors would not be set properly the next time
Stata was launched if the user changed the system appearance while the
Viewer was visible. This has been fixed.