Claude/mode submode display opinion xb193h - #1125
Merged
Merged
Conversation
Ham operators think of KLog's submode values (FT4, FT8, PSK31...) as the mode itself, so the log grid now labels that column "Mode". The true ADIF parent mode column (MFSK, PSK, RTTY...) is relabeled "Mode ADIF" to make clear it's the underlying ADIF classification rather than a second, unfamiliar concept.
Default log view now shows only the flattened "Mode" column (submode) out of the box; "Mode ADIF" (modeid) must be enabled explicitly via the column selector. Aligned the dead-code fallback default list in SetupPageLogView::init() with the real default set.
- Default column order (log grid, when no custom order is saved) now places modeid and submode immediately after bandid instead of following raw DB schema order. - Settings field list (SetupPageLogView) is reordered the same way, so Mode ADIF/Mode appear right after Band there too, and its "submode" entry is labeled "Mode (ADIF submode)" for clarity while the log view itself keeps the plain "Mode" label. - SearchWindow (DX/awards search) now shows both the real ADIF mode (Mode ADIF) and the flattened submode (Mode) as separate columns, instead of only the flattened value under "Mode". The award-status lookup, which relied on that flattened text, now reads it from the submode column instead of modeid, preserving its behaviour.
By default SearchWindow now shows only "Mode" (the flattened submode), positioned right after Band -- where "Mode ADIF" used to sit. "Mode ADIF" is only added, right after Band and before Mode, when the user has "modeid" active in their log view field settings, mirroring the main log view's own default-hidden/opt-in behaviour for that column.
modeid is the ADIF parent mode (e.g. MFSK) and is shared by every submode of that family, so keying the log table's uniqueness constraint on it collapsed genuinely different QSOs (e.g. FT2 vs FT4, both MFSK) that share callsign/band/timestamp/log. The second INSERT silently failed and was swallowed as an "expected duplicate", losing the QSO with no warning to the user. Adds DB migration step 0.031 (updateTo031 -> recreateTableLog) to apply the fixed constraint to existing databases. The migration is safe: any two rows that would collide on submode already shared the same modeid, so they could never have coexisted under the old constraint in the first place. Fixes #1119
modeid is the ADIF parent mode (e.g. MFSK) and is shared by every submode of that family, so the "This QSO seems to be duplicated" warning and the in-memory duplicate cache could flag genuinely different QSOs (e.g. FT2 vs FT4, both MFSK) as duplicates. Fixed consistently on both sides of the cache: loadDupeCacheBG() now builds it from log.submode instead of log.modeid, and every writer/ reader (isThisQSODuplicated, FileManager::processQSO, MainWindow::actionsJustAfterAddingOneQSO) now computes the mode id via the existing getSubModeIdFromQSO() helper instead of getIdFromModeName on the parent mode. filemanager.cpp previously had this backed out (see removed comment) because only some call sites had been switched to submode while the cache itself still keyed on modeid -- this fix moves everything over together so cache writes and reads stay consistent. Also simplified the unused QSO-based findDuplicateId() overload in dataproxy_sqlite.h to call getSubModeIdFromQSO() instead of duplicating its logic inline. Fixes #1120
logView never had an item delegate assigned, unlike every other relational table view in the app (logs, subdivisions, sats, world editor setup pages). Without QSqlRelationalDelegate, editing a relational column (band, mode ADIF, mode, country...) inline uses the default delegate, which edits/writes the raw foreign key id instead of offering a combobox of the related values, so the edit fails and the view reverts to the previous value.
Added LogModeDelegate (QSqlRelationalDelegate subclass) as the log grid's item delegate for the modeid/submode columns: - The Mode combobox only lists the user's active submodes (mirroring MainQSOEntryWidget's mode combo) instead of every submode in the DB. - The Mode ADIF combobox lists the distinct parent modes derived from those active submodes. - Picking a submode updates Mode ADIF to that submode's parent. - Picking a parent mode resets the submode to the mode-table row that stands for "just this mode" (submode field equal to its own name). - The cell's current value is always included even if since disabled, so opening the editor never silently changes it. LogWindow::setActiveModes() feeds the delegate the same active-modes list already pushed to MainQSOEntryWidget/MapWindow, wired in at both call sites in mainwindow.cpp.
…edit LogModeDelegate's combobox now commits and closes as soon as the user picks an item (wired to QComboBox::activated), instead of waiting for the editor to lose focus. Without this, the Mode ADIF/Mode sync only took effect once the user clicked elsewhere, reading as if the pick hadn't worked. Also removed DoubleClicked from logView's edit triggers. Once the log grid got a working delegate, double-clicking a cell both opened the QSO edit dialog (slotDoubleClickLog) and started inline editing on the same click, and the two fought over the double-click -- the dialog stopped opening. Double-click is now reserved for the edit dialog; inline editing is still reachable via F2/Enter or typing on a selected cell.
Added QAbstractItemView::SelectedClicked to logView's edit triggers, so the click gesture on the log grid is: 1st click selects the row (default), a further click on an already-selected row edits that cell, and double-click still opens the QSO edit dialog (unaffected, since DoubleClicked was already excluded from the edit triggers).
Replaced the built-in SelectedClicked edit trigger with manual handling: a click on the row the previous click already landed on arms a QTimer (system doubleClickInterval()) instead of starting the editor immediately. slotDoubleClickLog() cancels that timer as soon as a double-click is recognized, so double-clicking an already-selected row reliably opens the QSO edit dialog instead of racing a cell editor open from the first of the two clicks. m_lastClickedRow/pending state is reset in createlogPanel() so a row index left over from a previous log/reload can't be mistaken for "already selected" in the freshly loaded one.
The log grid selects whole rows (needed for the bulk QSO actions), so there was no visible way to tell which cell was "current" -- the one arrow keys move, and the one F2/click-on-already-selected-row would edit. LogModeDelegate now draws a border around the cell that has QStyle::State_HasFocus, in the palette's highlight color, on top of the normal row selection background, without changing the selection behaviour itself.
… grid" This reverts commit 803fa77.
…close getBandsInLog/getModesInLog/getSubModesInLog (DISTINCT/GROUP BY queries run every time the Settings dialog opens and again when it closes, via SetupDialog::readActiveBands/readActiveModes and MainWindow's own copies) and the duplicate-QSO lookups were full table scans: bandid, modeid, submode and lognumber had no index anywhere in the schema. Added DataBase::createLogIndexes(), called right after a fresh "log" table is created and again by recreateTableLog() once "logtemp" has been renamed back to "log". Indexing only the final "log" table (never the short-lived "logtemp" scratch table) and always under the same fixed name avoids two problems a naive per-call-site index would hit: slowing down logtemp's bulk INSERT for no benefit, and colliding on "index already exists" the next time some future migration calls recreateTableLog() again. Folded into updateTo031() (issue #1119), which already rebuilds the log table on upgrade -- no separate DB version bump needed.
DXClusterAssistant::modeFamily() collapsed every mode to its ADIF parent (e.g. MFSK) before comparing a logged QSO's mode against a spot's mode. FT2 and FT4 are both MFSK but are operationally different digital modes, so logging a QSO on FT4 caused removeSpotsOfLoggedQSO() to also remove -- and alreadyWorked() to treat as worked -- a still-workable FT2 spot for the same call/band. modeFamily() now compares by the actual submode instead, keeping only the one grouping that is genuinely correct: USB/LSB/SSB voice sidebands, which fold to a common "SSB" bucket as before. An unknown mode still returns empty, taking no part in the comparison, same as before. Fixes #1123
Awards::updateDXCCStatus() built dxccStatusList keyed by log.modeid (the ADIF parent mode, e.g. MFSK), while every caller of getQSOStatus() already computes a submode-specific mode id (currentModeShown, mainQSOEntryWidget's mode combo, SearchWindow, DXAssistantEngine all resolve to specific submodes like FT4). Since a submode id almost never equals the parent id stored in EntityStatus, the direct comparison in getQSOStatus() always failed: DXCC status showed "needed" even when already confirmed on that exact submode. Fixed by grouping/keying updateDXCCStatus()'s query and extractEntityStatus() by log.submode instead of log.modeid, matching what every caller already expects. DXCCStatusWidget worked around the mismatch by expanding to getModeGroupIds() (every mode sharing the same ADIF parent), which traded "never matches" for "matches too broadly": selecting FT4 showed the combined status of the whole MFSK family (FT2+FT4+FST4+JS8+Q65...). Added DataProxy_SQLite::getSidebandGroupIds(), which only groups the one family where that is actually intentional -- USB/LSB/SSB -- and switched DXCCStatusWidget to it. AwardsWidget's own getModeGroupIds() usage (counters, not status) is left for #1122, since it needs the same modeid->submode fix at the SQL level (buildModeInClause/getFieldInBand) to be worth changing. Fixes #1121
LogWindow::saveColumnOrder()/retoreColumsOrder() persisted the log grid's custom column order via QVariant::fromValue<QList<int>>, but QList<int> is not a type QVariant knows how to serialize/deserialize on its own. This produced the warning on every startup, and silently never restored a saved custom order -- the size-mismatch fallback in retoreColumsOrder() absorbed the failed load every time. Switched to storing it as a QStringList of numbers, the same pattern already used right below for ColumnWidths (a genuinely simple type QSettings/QVariant round-trip natively, no custom registration needed). Pre-existing, unrelated to this session's earlier changes; found via a user startup log.
buildModeInClause() (shared by getQSOonYear, getDXCConYear, getCQzonYear, getHowManyQSOInLog, getHowManyConfirmedQSLInLog), Awards::getWAZWorked/getWAZConfirmed, and the modeIds-list branch of getFieldInBand all filtered "AND modeid IN (...)" -- modeid is the ADIF parent mode (e.g. MFSK), shared by every submode of that family, so selecting FT4 in the Awards tab actually counted every MFSK submode (FT2, FST4, JS8, Q65...) too. getFieldInBand's other branch (the one taking _mode as a string) was already fixed to filter on submode via getSubModeFilterSQL -- this finishes that migration for the modeIds-list branch and its siblings. AwardsWidget::showAwards()/showDXMarathon() fed all of the above from dataProxy->getModeGroupIds(currentMode), which (independently of the modeid/submode column choice) over-broadens by grouping on shared ADIF parent -- the same bug already fixed for DXCCStatusWidget in #1121. Switched both to getSidebandGroupIds(), which only groups the one family where that's actually correct: USB/LSB/SSB. Fixes #1122
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.