| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
|
|
|
| |
TablesList's context and gear menu to avoid having two separators one after another
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
functionality, including an overhaul of filename NSTokenField functionality:
- Improve the NSTokenField used for export filenames: only tokenise reserved tokens, don't tokenise reserved words which are parts of other words, allow the comma to be used, update tokenisation during typing, and prevent whitespace triming.
- Save the last selected export path, and make the path selection button open a dialog to the selected directory
- Save the export filename and restore on future uses of the export dialog (only if the name contains placeholder tokens, so one-off export names aren't saved)
- If the advanced options are collapsed, display a summary of the selected options next to the disclosure triangle
- Display a small warning in the corner of the window if the export file cannot be imported into Sequel Pro, to warn those people attempting to back up their databases in XML
- Clarify and improve the export warning dialog if files already exist or could not be created; make the simpler file-exists cases reflect OS-style dialogs, alter wording based on the number of files that failed and how they failed, and only show the "replace" or "skip" type buttons if it makes sense to do so.
- Fix a mutation-during-enumeration error when skipping files
- If "Cancel" is chosen in the export file creation replace/error dialog, redisplay the export sheet with the previous selection still active
- Add support for year, month and day tokens in the filename token list
- Don't allow blank custom filenames, before or after tokenisation, as this can cause problems - instead fall back to default filenames in those cases
- Only append the extension if one hasn't been set - on all export formats, extending r3284
- If exporting to multiple files option is enabled but only one table is selected, supply that table name for filename table tokens
- Update the progress bar to reflect update progress when exporting CSV data
- Fix a bug causing exports to hang if the low-memory advanced option was set and content was selected to export and any empty tables were encountered
- Save memory use and compression advanced export settings across sessions
- Update localisable strings
|
| |
|
|
|
|
|
|
|
|
| |
- caught some more exceptional cases
- re-factored byte to bit conversion (really do not know why the old code wasn't working anymore); this should fix i1042
- disabled unisgned/zerofill check boxes in Structure for BIT
- insert correct default BIT field default value if any after adding a new row
-- this works for 5.5 but for MySQL 5.0.x it seems that providing a default value is a MySQL bug (further checks are needed)
|
|
|
|
| |
- fixes issue 1041
|
| |
|
|
|
|
|
|
| |
hidden pref key
* Change some [[NSPopupButton selectedItem] title] to [NSPopupButton titleOfSelectedItem]
|
|
|
|
| |
Fixes issue #1035.
|
| |
|
|
|
|
| |
add new rows/records
|
|
|
|
| |
their controller instead of SPPreferencesController; this fixes issues such as the favourite name not being generated correctly, or the password being lost when a username or database field is changed.
|
|
|
|
|
|
|
|
|
|
| |
null-terminated string processing:
- This fixes issues caused by null characters in strings - addressing Issue 1029
- Also appears to be a few percent faster than the old approach when processing lots of short strings
- Allows significant simplification of MCPResult and low-memory MCPStreamingResult code, avoiding a memory copy; this also gives a significant speedup and can actually make full streaming in MCPStreamingResult faster than "fast streaming".
The code will be reviewed further in future to improve on the gains seen here.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
postNotificationOnMainThread: which also exists in TCMPortMapper.framework; NSUserDefaults access now permitted in SP_REFACTOR blocks for a few prefs keys
|
|
|
|
|
|
| |
- Fix a few encoding issues
- Recode a few [NSApp keyWindow] uses back to [SPDatabaseDocument parentWindow] following discussion; possibly for later refactoring into use of an error method
|
|
|
|
| |
UI code from functional code
|
|
|
|
| |
HTML output window and bundle and field editors, adding commands for the equivalent functionality. This standardises text views with the custom query view and other applications
|
|
|
|
| |
setup of the table directly after creation
|
|
|
|
| |
this fixes Issue #1024
|
|
|
|
|
| |
- Use this new support in SPTablesList to prevent the tab key entering rename mode when focus is on the table list - this addresses Issue #1028
|
|
|
|
| |
keychain names and accounts. This addresses incorrect password use, fixing Issue #1020
|
|
|
|
|
|
|
| |
- This may be due to an Apple bug. I tracked it down to purely SPDatabaseDocument having a reference outlet to the activity scroll view.
- Removing the reference worked as a fix; removing all code that used that reference did not.
- Renaming the reference, and removing SPTableInfo's use of a reference, appears to act as a workaround. If the reference is not renamed, the bug still mainfests (!!!)
|
|
|
|
| |
Issue #1011
|
|
|
|
| |
favourites list
|
| |
|
|
|
|
|
|
|
| |
first-item preservation dates back to when the title had to be included in the favorites array, so removed
- Fix favorites sorting and preservation of unsorted status
|
|
|
|
|
|
|
| |
- Fix potential reload loop when setTableDetails: deselects the table row, after the fields have been emptied but before any edits have been committed
- Fix reload action not checking whether items need saving before comparisons occur
- Save incomplete edits back to the table before processing changes to avoid changes requiring additional queries
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
passing
|
| |
|
|
|
|
|
| |
- Tweak README
|
| |
|
| |
|
| |
|
|
|
|
|
| |
- fixes i996
- discussion: why do we need [NSWindow endEditingFor:] ??
|
|
|
|
| |
it's 10.6 method (but it worked?)
|
| |
|
| |
|
|
|
|
| |
- due to http://dev.mysql.com/doc/refman/5.1/en/mysql-affected-rows.html mysql_affected_rows() normally returns a my_ulonglong BUT it can return -1; thus we have to check whether it is >= 0 via [mySQLConnection affectedRows] != (my_ulonglong)~0
|
| |
|
| |
|
| |
|