verdnatura-chat/ios/Pods/Flipper-DoubleConversion/double-conversion/fast-dtoa.cc

666 lines
30 KiB
C++
Raw Normal View History

Merge beta into master (#2143) * [FIX] Messages being sent but showing as temp status (#1469) * [FIX] Missing messages after reconnect (#1470) * [FIX] Few fixes on themes (#1477) * [I18N] Missing German translations (#1465) * Missing German translation * adding a missing space behind colon * added a missing space after colon * and another attempt to finally fix this – got confused by all the branches * some smaller fixes for the translation * better wording * fixed another typo * [FIX] Crash while displaying the attached image with http on file name (#1401) * [IMPROVEMENT] Tap app and server version to copy to clipboard (#1425) * [NEW] Reply notification (#1448) * [FIX] Incorrect background color login on iPad (#1480) * [FIX] Prevent multiple tap on send (Share Extension) (#1481) * [NEW] Image Viewer (#1479) * [DOCS] Update Readme (#1485) * [FIX] Jitsi with Hermes Enabled (#1523) * [FIX] Draft messages not working with themed Messagebox (#1525) * [FIX] Go to direct message from members list (#1519) * [FIX] Make SAML wait for idp token instead of creating it on client (#1527) * [FIX] Server Test Push Notification (#1508) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [CHORE] Update to new server response (#1509) * [FIX] Insert messages with blank users (#1529) * Bump version to 4.2.1 (#1530) * [FIX] Error when normalizing empty messages (#1532) * [REGRESSION] CAS (#1570) * Bump version to 4.2.2 (#1571) * [FIX] Add username block condition to prevent error (#1585) * Bump version to 4.2.3 * Bump version to 4.2.4 * Bump version to 4.3.0 (#1630) * [FIX] Channels doesn't load (#1586) * [FIX] Channels doesn't load * [FIX] Update roomsUpdatedAt when subscriptions.length is 0 * [FIX] Remove unnecessary changes * [FIX] Improve the code Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Make SAML to work on Rocket.Chat < 2.3.0 (#1629) * [NEW] Invite links (#1534) * [FIX] Set the http-agent to the form that Rocket.Chat requires for logging (#1482) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] "Following thread" and "Unfollowed Thread" is hardcoded and not translated (#1625) * [FIX] Disable reset button if form didn't changed (#1569) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Header title of RoomInfoView (#1553) * [I18N] Gallery Permissions DE (#1542) * [FIX] Not allow to send messages to archived room (#1623) * [FIX] Profile fields automatically reset (#1502) * [FIX] Show attachment on ThreadMessagesView (#1493) * [NEW] Wordpress auth (#1633) * [CHORE] Add Start Packager script (#1639) * [CHORE] Update RN to 0.61.5 (#1638) * [CHORE] Update RN to 0.61.5 * [CHORE] Update react-native patch Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * Bump version to 4.3.1 (#1641) * [FIX] Change force logout rule (#1640) * Bump version to 4.4.0 (#1643) * [IMPROVEMENT] Use MessagingStyle on Android Notification (#1575) * [NEW] Request review (#1627) * [NEW] Pull to refresh RoomView (#1657) * [FIX] Unsubscribe from room (#1655) * [FIX] Server with subdirs (#1646) * [NEW] Clear cache (#1660) * [IMPROVEMENT] Memoize and batch subscriptions updates (#1642) * [FIX] Disallow empty sharing (#1664) * [REGRESSION] Use HTTPS links for sharing and markets protocol for review (#1663) * [FIX] In some cases, share extension doesn't load images (#1649) * [i18n] DE translations for new invite function and some minor fixes (#1631) * [FIX] Remove duplicate jetify step (#1628) minor: also remove 'cd' calls Co-authored-by: Diego Mello <diegolmello@gmail.com> * [REGRESSION] Read messages (#1666) * [i18n] German translations missing (#1670) * [FIX] Notifications crash on older Android Versions (#1672) * [i18n] Added Dutch translation (#1676) * [NEW] Omnichannel Beta (#1674) * [NEW] Confirm logout/clear cache (#1688) * [I18N] Add es-ES language (#1495) * [NEW] UiKit Beta (#1497) * [IMPROVEMENT] Use reselect (#1696) * [FIX] Notification in Android API level less than 24 (#1692) * [IMPROVEMENT] Send tmid on slash commands and media (#1698) * [FIX] Unhandled action on UIKit (#1703) * [NEW] Pull to refresh RoomsList (#1701) * [IMPROVEMENT] Reset app when language is changed (#1702) * [FIX] Small fixes on UIKit (#1709) * [FIX] Spotlight (#1719) * [CHORE] Update react-native-image-crop-picker (#1712) * [FIX] Messages Overlapping (Android) and MessageBox Scroll (iOS) (#1720) * [REGRESSION] Remove @ and # from mention (#1721) * [NEW] Direct message from user info (#1516) * [FIX] Delete slash commands (#1723) * [IMPROVEMENT] Hold URL to copy (#1684) * [FIX] Different sourcemaps generation for Hermes (#1724) * [FIX] Different sourcemaps generation for Hermes * Upload sourcemaps after build * [REVERT] Show emoji keyboard on Android (#1738) * [FIX] Stop logging react-native-image-crop-picker (#1745) * [FIX] Prevent toast ref error (#1744) * [FIX] Prevent reaction map error (#1743) * [FIX] Add missing calls to user info (#1741) * [FIX] Catch room unsubscribe error (#1739) * [i18n] Missing German keys (#1735) * [FIX] Missing i18n on MessagesView title (#1733) * [FIX] UIKit Modal: Weird behavior on Android Tablet (#1742) * [i18n] Missing key on German (#1747) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [i18n] Add Italian (#1736) * [CHORE] Memory leaks investigation (#1675) * [IMPROVEMENT] Alert verify email when enabled (#1725) * [NEW] Jitsi JWT added to URL (#1746) * [FIX] UIKit submit when connection lost (#1748) * Bump version to 4.5.0 (#1761) * [NEW] Default browser (#1752) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] HTTP Basic Auth (#1753) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Honor profile fields edit settings (#1687) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Room announcements (#1726) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Honor Register/Login settings (#1727) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Make links clickable on Room Info (#1730) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [NEW] Hide system messages (#1755) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Honor "Message_AudioRecorderEnabled" (#1764) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [i18n] Missing de keys (#1765) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Redirect user to SetUsernameView (#1728) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Join Room (#1769) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Accept all media types using * (#1770) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Use RealName when necessary (#1758) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Markdown Line Break (#1783) * [IMPROVEMENT] Remove useMarkdown (#1774) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Open browser rather than webview on Create Workspace (#1788) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Markdown perf (#1796) * [FIX] Stop video when modal is closed (#1787) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Hide reply notification action when there are missing data (#1771) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [i18n] Added Japanese translation (#1781) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Reset password error message (#1772) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Close tablet modal (#1773) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Setting not present (#1775) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Thread header (#1776) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Keyboard tracking loses input ref (#1784) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [NEW] Mark message as unread (#1785) Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * [IMPROVEMENT] Log server version (#1786) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Add loading message on long running tasks (#1798) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [CHORE] Switch Apple account on Fastlane (#1810) * [FIX] Watermelon throwing "Cannot update a record with pending updates" (#1754) * [FIX] Detox tests (#1790) * [CHORE] Use markdown preview on RoomView Header (#1807) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] LoginSignup blink services (#1809) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Request user presence on demand (#1813) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Remove all invited users when create a channel (#1814) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Pop from room which you have been removed (#1819) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Room Info styles (#1820) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [i18n] Add missing German keys (#1800) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Empty mentions for @all and @here when real name is enabled (#1822) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [TESTS] Markdown added to Storybook (#1812) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [REGRESSION] Room View header title (#1827) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Storybook snapshots (#1831) Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * [FIX] Mentions (#1829) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Thread message not found (#1830) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Separate delete and remove channel (#1832) * Rename to delete room * Separate delete and remove channel * handleRemoved -> handleRoomRemoved * [FIX] Navigate to RoomsList & Handle tablet case Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * [NEW] Filter system messages per room (#1815) Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] e2e tests (#1838) * [FIX] Consecutive clear cache calls freezing app (#1851) * Bump version to 4.5.1 (#1853) * [FIX][iOS] Ignore silent mode on audio player (#1862) * [IMPROVEMENT] Create App Group property on Info.plist (#1858) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Make username clickable on message (#1618) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Show proper error message on profile (#1768) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Show toast when a message is starred/unstarred (#1616) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Incorrect size params to avatar endpoint (#1875) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Remove unrecognized emoji flags on android (#1887) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Remove react-native global installs (#1886) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Emojis transparent on android (#1881) Co-authored-by: Diego Mello <diegolmello@gmail.com> * Bump acorn from 5.7.3 to 5.7.4 (#1876) Bumps [acorn](https://github.com/acornjs/acorn) from 5.7.3 to 5.7.4. - [Release notes](https://github.com/acornjs/acorn/releases) - [Commits](https://github.com/acornjs/acorn/compare/5.7.3...5.7.4) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * Bump version to 4.6.0 (#1911) * [FIX] Encode Image URI (#1909) * [FIX] Encode Image URI * [FIX] Check if Image is Valid Co-authored-by: Diego Mello <diegolmello@gmail.com> * [NEW] Adaptive Icons (#1904) * Remove unnecessary stuff from debug build * Adaptive icon for experimental app * [FIX] Stop showing message on leave channel (#1896) * [FIX] Leave room don't show 'was removed' message * [FIX] Remove duplicated code Co-authored-by: Diego Mello <diegolmello@gmail.com> * [i18n] Added missing German translations(#1900) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Linkedin OAuth login (#1913) * [CHORE] Fix typo in CreateChannel View (#1930) * [FIX] Respect protocol in HTTP Auth IPs (#1933) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Use new LinkedIn OAuth url (#1935) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [CHORE] Use storyboard on splash screen (#1939) * Update react-native-bootsplash * iOS * Fix android * [FIX] Check if avatar exists before create Icon (#1927) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Ignore self typing event (#1950) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Change default directory listing to Users (#1948) * fix: change default directory listing to Users * follow server settings * Fix state to props Co-authored-by: Diego Mello <diegolmello@gmail.com> * [NEW] Onboarding layout (#1954) * Onboarding texts * OnboardingView * FormContainer * Minor fixes * NewServerView * Remove code * Refactor * WorkspaceView * Stash * Login with email working * Login with * Join open * Revert "Login with" This reverts commit d05dc507d2e9a2db76d433b9b1f62192eba35dbd. * Fix create account styles * Register * Refactor * LoginServices component * Refactor * Multiple servers * Remove native images * Refactor styles * Fix testid * Fix add server on tablet * i18n * Fix close modal * Fix TOTP * [FIX] Registration disabled * [FIX] Login Services separator * Fix logos * Fix AppVersion name * I18n * Minor fixes * [FIX] Custom Fields Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * [NEW] Create discussions (#1942) * [WIP][NEW] Create Discussion * [FIX] Clear multiselect & Translations * [NEW] Create Discussion at MessageActions * [NEW] Disabled Multiselect * [FIX] Initial channel * [NEW] Create discussion on MessageBox Actions * [FIX] Crashing on edit name * [IMPROVEMENT] New message layout * [CHORE] Update README * [NEW] Avatars on MultiSelect * [FIX] Select Users * [FIX] Add redirect and Handle tablet * [IMPROVEMENT] Split CreateDiscussionView * [FIX] Create a discussion inner discussion * [FIX] Create a discussion * [I18N] Add pt-br * Change icons * [FIX] Nav to discussion & header title * Fix header Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Load messages (#1910) * Create updateLastOpen param on readMessages * Remove InteractionManager from load messages * [NEW] Custom Status (#1811) * [NEW] Custom Status * [FIX] Subscribe to changes * [FIX] Improve code using Banner component * [IMPROVEMENT] Toggle modal * [NEW] Edit custom status from Sidebar * [FIX] Modal when tablet * [FIX] Styles * [FIX] Switch to react-native-promp-android * [FIX] Custom Status UI * [TESTS] E2E Custom Status * Fix banner * Fix banner * Fix subtitle * status text * Fix topic header * Fix RoomActionsView topic * Fix header alignment on Android * [FIX] RoomInfo crashes when without statusText * [FIX] Use users.setStatus * [FIX] Remove customStatus of ProfileView * [FIX] Room View Thread Header Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] UI issues of Create Discussion View (#1965) * [NEW] Direct Message between multiple users (#1958) * [WIP] DM between multiple users * [WIP][NEW] Create new DM between multiple users * [IMPROVEMENT] Improve createChannel Sagas * [IMPROVEMENT] Selected Users view * [IMPROVEMENT] Room Actions of Group DM * [NEW] Create new DM between multiple users * [NEW] Group DM avatar * [FIX] Directory border * [IMPROVEMENT] Use isGroupChat * [CHORE] Remove legacy getRoomMemberId * [NEW] RoomTypeIcon * [FIX] No use legacy method on RoomInfoView * [FIX] Blink header when create new DM * [FIX] Only show create direct message option when allowed * [FIX] RoomInfoView * pt-BR * Few fixes * Create button name * Show create button only after a user is selected * Fix max users issues Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Add server and hide login (#1968) * Navigate to new server workspace from ServerDropdown if there's no token * Hide login button based on login services and Accounts_ShowFormLogin setting * [FIX] Lint Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * [FIX] MultiSelect Keyboard behavior (Android) (#1969) * fixed-modal-position * made-changes Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * [FIX] Bottom border style on DirectoryView (#1963) * [FIX] Border style * [FIX] Refactoring * [FIX] fix color of border * Undo Co-authored-by: Aroo <azhaubassar@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Clear settings on server change (#1967) * [FIX] Deeplinking without RoomId (#1925) * [FIX] Deeplinking without rid * [FIX] Join channel * [FIX] Deep linking without rid * Update app/lib/methods/canOpenRoom.js Co-authored-by: Diego Mello <diegolmello@gmail.com> * [NEW] Two Factor authentication via email (#1961) * First api call working * [NEW] REST API Post wrapper 2FA * [NEW] Send 2FA on Email * [I18n] Add translations * [NEW] Translations & Cancel totp * [CHORE] Totp -> TwoFactor * [NEW] Two Factor by email * [NEW] Tablet Support * [FIX] Text colors * [NEW] Password 2fa * [FIX] Encrypt password on 2FA * [NEW] MethodCall2FA * [FIX] Password fallback * [FIX] Wrap all post/methodCall with 2fa * [FIX] Wrap missed function * few fixes * [FIX] Use new TOTP on Login * [improvement] 2fa methodCall Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * [FIX] Correct message for manual approval user Registration (#1906) * [FIX] Correct message for manual approval from admin shown on Registeration * lint fix - added semicolon * Updated the translations * [FIX] Translations * i18n to match server Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Direct Message between multiple users REST (#1974) * [FIX] Investigate app losing connection issues (#1890) * [WIP] Reopen without timeOut & ping with 5 sec & Fix Unsubscribe * [FIX] Remove duplicated close * [FIX] Use no-dist lib * [FIX] Try minor fix * [FIX] Try reopen connection when app was put on foreground * [FIX] Remove timeout * [FIX] Build * [FIX] Patch * [FIX] Snapshot * [IMPROVEMENT] Decrease time to reopen * [FIX] Some fixes * [FIX] Update sdk version * [FIX] Subscribe Room Once * [CHORE] Update sdk * [FIX] Subscribe Room * [FIX] Try to resend missed subs * [FIX] Users never show status when start app without network * [FIX] Subscribe to room * [FIX] Multiple servers * [CHORE] Update SDK * [FIX] Don't duplicate streams on subscribeAll * [FIX] Server version when start the app offline * [FIX] Server version cached * [CHORE] Remove unnecessary code * [FIX] Offline server version * [FIX] Subscribe before connect * [FIX] Remove unncessary props * [FIX] Update sdk * [FIX] User status & Unsubscribe Typing * [FIX] Typing at incorrect room * [FIX] Multiple Servers * [CHORE] Update SDK * [REVERT] Undo some changes on SDK * [CHORE] Update sdk to prevent incorrect subscribes * [FIX] Prevent no reconnect * [FIX] Remove close on open * [FIX] Clear typing when disconnect/connect to SDK * [CHORE] Update SDK * [CHORE] Update SDK * Update SDK * fix merge develop Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Single message thread inserting thread without rid (#1999) * [FIX] ThreadMessagesView crashing on load (#1997) * [FIX] Saml (#1996) * [FIX] SAML incorrect close * [FIX] Pathname Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Change user own status (#1995) * [FIX] Change user own status * [IMPROVEMENT] Set activeUsers Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Loading all updated rooms after app resume (#1998) * [FIX] Loading all updated rooms after app resume * Fix room date on RoomItem Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Change notifications preferences (#2000) * [FIX] Change notifications preferences * [IMPROVEMENT] Picker View * [I18N] Translations * [FIX] Picker Selection * [FIX] List border * [FIX] Prevent crash * [FIX] Not-Pref tablet * [FIX] Use same style of LanguageView * [IMPROVEMENT] Send listItem title Co-authored-by: Diego Mello <diegolmello@gmail.com> * Bump version to 4.6.1 (#2001) * [FIX] DM header blink (#2011) * [FIX] Split get settings into two requests (#2017) * [FIX] Split get settings into two requests * [FIX] Clear settings only when change server * [IMPROVEMENT] Move the way to clear settings * [REVERT] Revert some changes * [FIX] Server Icon Co-authored-by: Diego Mello <diegolmello@gmail.com> * [REGRESSION] Invite Links (#2007) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Read only channel/broadcast (#1951) * [FIX] Read only channel/broadcast * [FIX] Roles missing * [FIX] Check roles to readOnly * [FIX] Can post * [FIX] Respect post-readonly permission * [FIX] Search a room readOnly Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Cas auth (#2024) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Login TOTP Compatibility to older servers (#2018) * [FIX] Login TOTP Compatibility to older servers * [FIX] Android crashes if use double negation Co-authored-by: Diego Mello <diegolmello@gmail.com> * Bump version to 4.6.4 (#2029) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Lint (#2030) * [FIX] UIKit with only one block (#2022) * [FIX] Message with only one block * [FIX] Update headers Co-authored-by: Diego Mello <diegolmello@gmail.com> * Bump version to 4.7.0 (#2035) * [FIX] Action Tint Color on Black theme (#2081) * [FIX] Prevent crash when thread is not found (#2080) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Prevent double click (#2079) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Show slash commands when disconnected (#2078) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Backhandler onboarding (#2077) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Respect UI_Allow_room_names_with_special_chars setting (#2076) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] RoomsList update sometimes isn't fired (#2071) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Stop inserting last message as message object from rooms stream if room is focused (#2069) * [IMPROVEMENT] No insert last message if the room is focused * fix discussion/threads Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Hide system messages (#2067) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Pending update (#2066) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Prevent crash when room.uids was not inserted yet (#2055) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FEATURE] Save video (#2063) * added-feature-save-video * fix sha256 Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Send totp-code to meteor call (#2050) * fixed-issue * removed-variable-name-errors * reverted-last-commit Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] MessageBox mention shouldn't show group DMs (#2049) * fixed-issue * [FIX] Filter users only if it's not a group chat Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] AttachmentView (Android)(Tablet) (#2047) * [fix]Tablet attachment View and Room Navigation * fix weird navigation and margin bottom Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Allow special chars in Filename (#2020) * fixed-filename-issue * improve Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Recorded audio on Android doesn't play on iOS (#2073) * react-native-video -> expo-av * remove react-native-video * Add audio mode * update mocks * [FIX] Loading bigger than play/pause Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Message Touchable (#2082) * [FIX] Avatar touchable * [IMPROVEMENT] onLongPress on all Message Touchables * [IMPROVEMENT] User & baseUrl on MessageContext * [FIX] Context Access * [FIX] BaseURL * Fix User Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] ReactionsModal (#2085) * [NEW] Delete Server (#1975) * [NEW] Delete server Co-authored-by: Bruno Dantas <oliveiradantas96@gmail.com> Co-authored-by: Calebe Rios <calebersmendes@gmail.com> * [FIX] Revert removed function Co-authored-by: Bruno Dantas <oliveiradantas96@gmail.com> Co-authored-by: Calebe Rios <calebersmendes@gmail.com> * pods * i18n * Revert "pods" This reverts commit 2854a1650538159aeeafe90fdb2118d12b76a82f. Co-authored-by: Bruno Dantas <oliveiradantas96@gmail.com> Co-authored-by: Calebe Rios <calebersmendes@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Change server while connecting/updating (#1981) * [IMPROVEMENT] Change server while connecting * [FIX] Not login/reconnect to previous server * [FIX] Abort all fetch while connecting * [FIX] Abort sdk fetch * [FIX] Patch-package * Add comments Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Keep screen awake while recording/playing some audio (#2089) * [IMPROVEMENT] Keep screen awake while recording/playing some audio * [FIX] Add expo-keep-awake mock * [FIX] UIKit crashing when UIKitModal receive update event (#2088) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Close announcement banner (#2064) * [NEW] Created new field in subscription table Signed-off-by: Ezequiel De Oliveira <ezequiel1de1oliveira@gmail.com> * [NEW] New field added to obeserver in room view Signed-off-by: Ezequiel De Oliveira <ezequiel1de1oliveira@gmail.com> * [NEW] Added icon and new design to banner Signed-off-by: Ezequiel De Oliveira <ezequiel1de1oliveira@gmail.com> * [NEW] Close banner function works Signed-off-by: Ezequiel De Oliveira <ezequiel1de1oliveira@gmail.com> * [IMPROVEMENT] closed banner status now update correctly Signed-off-by: Ezequiel De Oliveira <ezequiel1de1oliveira@gmail.com> * improve banner style Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * Update all dependencies (#2008) * Android RN 62 * First steps iOS * Second step iOS * iOS compiling * "New" build system * Finish iOS * Flipper * Update to RN 0.62.1 * expo libs * Hermes working * Fix lint * Fix android build * Patches * Dev patches * Patch WatermelonDB: https://github.com/Nozbe/WatermelonDB/pull/660 * Fix jitsi * Update several minors * Update dev minors and lint * react-native-keyboard-input * Few updates * device info * react-native-fast-image * Navigation bar color * react-native-picker-select * webview * reactotron-react-native * Watermelondb * RN 0.62.2 * Few updates * Fix selection * update gems * remove lib * finishing * tests * Use node 10 * Re-enable app bundle * iOS build * Update jitsi ios * [NEW] Passcode and biometric unlock (#2059) * Update expo libs * Configure expo-local-authentication * ScreenLockedView * Authenticate server change * Auth on app resume * localAuthentication util * Add servers.lastLocalAuthenticatedSession column * Save last session date on background * Use our own version of app state redux * Fix libs * Remove inactive * ScreenLockConfigView * Apply on saved data * Auto lock option label * Starting passcode * Basic passcode flow working * Change passcode * Check if biometry is enrolled * Use fork * Migration * Patch expo-local-authentication * Use async storage * Styling * Timer * Refactor * Lock orientation portrait when not on tablet * share extension * Deep linking * Share extension * Refactoring passcode * use state * Stash * Refactor * Change passcode * Animate dots on error * Matching passcodes * Shake * Remove lib * Delete button * Fade animation on modal * Refactoring * ItemInfo * I18n * I18n * Remove unnecessary prop * Save biometry column * Raise time to lock to 30 seconds * Vibrate on wrong confirmation passcode * Reset attempts and save last authentication on local passcode confirmation * Remove inline style * Save last auth * Fix header blink * Change function name * Fix android modal * Fix vibration permission * PasscodeEnter calls biometry * Passcode on the state * Biometry button on PasscodeEnter * Show whole passcode * Secure passcode * Save passcode with promise to prevent empty passcodes and immediately lock * Patch expo-local-authentication * I18n * Fix biometry being called every time * Blur screen on app inactive * Revert "Blur screen on app inactive" This reverts commit a4ce812934adcf6cf87eb1a92aec9283e2f26753. * Remove immediately because of how Activities work on Android * Pods * New layout * stash * Layout refactored * Fix icons * Force set passcode from server * Lint * Improve permission message * Forced passcode subtitle * Disable based on admin's choice * Require local authentication on login success * Refactor * Update tests * Update react-native-device-info to fix notch * Lint * Fix modal * Fix icons * Fix min auto lock time * Review * keep enabled on mobile * fix forced by admin when enable unlock with passcode * use DEFAULT_AUTO_LOCK when manual enable screenLock * fix check has passcode * request biometry on first password * reset auto time lock when disabled on server Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> * [FIX] Messages View (#2090) * [FIX] Messages View * [FIX] Opening PDF from Files View * [FIX] Audio * [FIX] SearchMessagesView Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Big names overflow (#2072) * [FIX] Big names overflow * [FIX] Message time Co-authored-by: devyaniChoubey <devyanichoubey16@gmail.com> * [FIX] Some alignments * fix user item overflow * some adjustments Co-authored-by: devyaniChoubey <devyanichoubey16@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Avatar of message as an emoji (#2038) * fixed-issue * removed-hardcoded-emoji * Merge develop * replaced markdown with emoji componenent * made-changes * use avatar onPress Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Diego Mello <diegolmello@gmail.com> * [NEW] Livechat (#2004) * [WIP][NEW] Livechat info/actions * [IMPROVEMENT] RoomActionsView * [NEW] Visitor Navigation * [NEW] Get Department REST * [FIX] Borders * [IMPROVEMENT] Refactor RoomInfo View * [FIX] Error while navigate from mention -> roomInfo * [NEW] Livechat Fields * [NEW] Close Livechat * [WIP] Forward livechat * [NEW] Return inquiry * [WIP] Comment when close livechat * [WIP] Improve roomInfo * [IMPROVEMENT] Forward room * [FIX] Department picker * [FIX] Picker without results * [FIX] Superfluous argument * [FIX] Check permissions on RoomActionsView * [FIX] Livechat permissions * [WIP] Show edit to livechat * [I18N] Add pt-br translations * [WIP] Livechat Info * [IMPROVEMENT] Livechat info * [WIP] Livechat Edit * [WIP] Livechat edit * [WIP] Livechat Edit * [WIP] Livechat edit scroll * [FIX] Edit customFields * [FIX] Clean livechat customField * [FIX] Visitor Navigation * [NEW] Next input logic LivechatEdit * [FIX] Add livechat data to subscription * [FIX] Revert change * [NEW] Livechat user Status * [WIP] Livechat tags * [NEW] Edit livechat tags * [FIX] Prevent some crashes * [FIX] Forward * [FIX] Return Livechat error * [FIX] Prevent livechat info crash * [IMPROVEMENT] Use input style on forward chat * OnboardingSeparator -> OrSeparator * [FIX] Go to next input * [NEW] Added some icons * [NEW] Livechat close * [NEW] Forward Room Action * [FIX] Livechat edit style * [FIX] Change status logic * [CHORE] Remove unnecessary logic * [CHORE] Remove unnecessary code * [CHORE] Remove unecessary case * [FIX] Superfluous argument * [IMPROVEMENT] Submit livechat edit * [CHORE] Remove textInput type * [FIX] Livechat edit * [FIX] Livechat Edit * [FIX] Use same effect * [IMPROVEMENT] Tags input * [FIX] Add empty tag * Fix minor issues * Fix typo * insert livechat room data to our room object * review * add method calls server version Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Delete Subs (#2091) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Android build (#2094) * [FIX] Blink header DM (#2093) * [FIX] Blink header DM * Remove query * [FIX] Push RoomInfoView * remove unnecessary try/catch * [FIX] RoomInfo > Message (Tablet) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Default biometry enabled (#2095) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [IMPROVEMENT] Enable navigating to a room from auth deep linking (#2115) * Wait for login success to navigate * Enable auth and room deep linking at the same time * [FIX] NewMessageView Press Item should open DM (#2116) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Roles throwing error (#2110) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Wait attach activity before changeNavigationBarColor (#2111) * [FIX] Wait attach activity before changeNavigationBarColor * Remove timeout and add try/catch Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] UIKit crash when some app send a list (#2117) * [FIX] StoryBook * [FIX] UIKit crash when some app send a list * [CHORE] Update snapshot * [CHORE] Remove token & id * [FIX] Change bar color while no activity attached (#2130) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Screen Lock options i18n (#2120) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [i18n] Added missing German translation strings (#2105) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Sometimes SDK is null when try to connect (#2131) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [FIX] Autocomplete position on Android (#2106) * [FIX] Autocomplete position on Android * [FIX] Set selection to 0 when needed Co-authored-by: Diego Mello <diegolmello@gmail.com> * Revert "[FIX] Autocomplete position on Android (#2106)" (#2136) This reverts commit e8c38d6f6f69ae396a4aae6e37336617da739a6d. * [FIX] Here and all mentions shouldn't refer to users (#2137) * [FIX] No send data to bugsnag if it's an aborted request (#2133) Co-authored-by: Diego Mello <diegolmello@gmail.com> * [TESTS] Update and separate E2E tests (#2126) * Tests passing until roomslist * create room * roominfo * change server * broadcast * profile * custom status * forgot password * working * room and onboarding * Tests separated * config.yml refactor * Revert "config.yml refactor" This reverts commit 0e984d3029e47612726bf199553f7abdf24843e5. * CI * lint * CI refactor * Onboarding tests * npx detox * Add all tests * Save brew cache * mac-env executor * detox-test command * Update readme * Remove folder * [FIX] Screen Lock Time respect local value (#2141) * [FIX] Screen Lock Time respect local value * [FIX] Enable biometry at the first passcode change Co-authored-by: phriedrich <info@phriedrich.de> Co-authored-by: Guilherme Siqueira <guilhersiqueira@gmail.com> Co-authored-by: Prateek Jain <44807945+Prateek93a@users.noreply.github.com> Co-authored-by: Djorkaeff Alexandre <djorkaeff.unb@gmail.com> Co-authored-by: Prateek Jain <prateek93a@gmail.com> Co-authored-by: devyaniChoubey <52153085+devyaniChoubey@users.noreply.github.com> Co-authored-by: Bernard Seow <ssbing99@gmail.com> Co-authored-by: Hiroki Ishiura <ishiura@ja2.so-net.ne.jp> Co-authored-by: Exordian <jakob.englisch@gmail.com> Co-authored-by: Daanchaam <daanhendriks97@gmail.com> Co-authored-by: Youssef Muhamad <emaildeyoussefmuhamad@gmail.com> Co-authored-by: Iván Álvarez <ialvarezpereira@gmail.com> Co-authored-by: Sarthak Pranesh <41206172+sarthakpranesh@users.noreply.github.com> Co-authored-by: Michele Pellegrini <pellettiero@users.noreply.github.com> Co-authored-by: Tanmoy Bhowmik <tanmoy.openroot@gmail.com> Co-authored-by: Hibikine Kage <14365761+hibikine@users.noreply.github.com> Co-authored-by: Ezequiel de Oliveira <ezequiel1de1oliveira@gmail.com> Co-authored-by: Neil Agarwal <neil@neilagarwal.me> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Govind Dixit <GOVINDDIXIT93@GMAIL.COM> Co-authored-by: Zhaubassarova Aruzhan <49000079+azhaubassar@users.noreply.github.com> Co-authored-by: Aroo <azhaubassar@gmail.com> Co-authored-by: Sarthak Pranesh <sarthak.pranesh2018@vitstudent.ac.in> Co-authored-by: Siddharth Padhi <padhisiddharth31@gmail.com> Co-authored-by: Bruno Dantas <oliveiradantas96@gmail.com> Co-authored-by: Calebe Rios <calebersmendes@gmail.com> Co-authored-by: devyaniChoubey <devyanichoubey16@gmail.com>
2020-05-25 20:54:27 +00:00
// Copyright 2012 the V8 project authors. All rights reserved.
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// * Redistributions of source code must retain the above copyright
// notice, this list of conditions and the following disclaimer.
// * Redistributions in binary form must reproduce the above
// copyright notice, this list of conditions and the following
// disclaimer in the documentation and/or other materials provided
// with the distribution.
// * Neither the name of Google Inc. nor the names of its
// contributors may be used to endorse or promote products derived
// from this software without specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#include "fast-dtoa.h"
#include "cached-powers.h"
#include "diy-fp.h"
#include "ieee.h"
namespace double_conversion {
// The minimal and maximal target exponent define the range of w's binary
// exponent, where 'w' is the result of multiplying the input by a cached power
// of ten.
//
// A different range might be chosen on a different platform, to optimize digit
// generation, but a smaller range requires more powers of ten to be cached.
static const int kMinimalTargetExponent = -60;
static const int kMaximalTargetExponent = -32;
// Adjusts the last digit of the generated number, and screens out generated
// solutions that may be inaccurate. A solution may be inaccurate if it is
// outside the safe interval, or if we cannot prove that it is closer to the
// input than a neighboring representation of the same length.
//
// Input: * buffer containing the digits of too_high / 10^kappa
// * the buffer's length
// * distance_too_high_w == (too_high - w).f() * unit
// * unsafe_interval == (too_high - too_low).f() * unit
// * rest = (too_high - buffer * 10^kappa).f() * unit
// * ten_kappa = 10^kappa * unit
// * unit = the common multiplier
// Output: returns true if the buffer is guaranteed to contain the closest
// representable number to the input.
// Modifies the generated digits in the buffer to approach (round towards) w.
static bool RoundWeed(Vector<char> buffer,
int length,
uint64_t distance_too_high_w,
uint64_t unsafe_interval,
uint64_t rest,
uint64_t ten_kappa,
uint64_t unit) {
uint64_t small_distance = distance_too_high_w - unit;
uint64_t big_distance = distance_too_high_w + unit;
// Let w_low = too_high - big_distance, and
// w_high = too_high - small_distance.
// Note: w_low < w < w_high
//
// The real w (* unit) must lie somewhere inside the interval
// ]w_low; w_high[ (often written as "(w_low; w_high)")
// Basically the buffer currently contains a number in the unsafe interval
// ]too_low; too_high[ with too_low < w < too_high
//
// too_high - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
// ^v 1 unit ^ ^ ^ ^
// boundary_high --------------------- . . . .
// ^v 1 unit . . . .
// - - - - - - - - - - - - - - - - - - - + - - + - - - - - - . .
// . . ^ . .
// . big_distance . . .
// . . . . rest
// small_distance . . . .
// v . . . .
// w_high - - - - - - - - - - - - - - - - - - . . . .
// ^v 1 unit . . . .
// w ---------------------------------------- . . . .
// ^v 1 unit v . . .
// w_low - - - - - - - - - - - - - - - - - - - - - . . .
// . . v
// buffer --------------------------------------------------+-------+--------
// . .
// safe_interval .
// v .
// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - .
// ^v 1 unit .
// boundary_low ------------------------- unsafe_interval
// ^v 1 unit v
// too_low - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
//
//
// Note that the value of buffer could lie anywhere inside the range too_low
// to too_high.
//
// boundary_low, boundary_high and w are approximations of the real boundaries
// and v (the input number). They are guaranteed to be precise up to one unit.
// In fact the error is guaranteed to be strictly less than one unit.
//
// Anything that lies outside the unsafe interval is guaranteed not to round
// to v when read again.
// Anything that lies inside the safe interval is guaranteed to round to v
// when read again.
// If the number inside the buffer lies inside the unsafe interval but not
// inside the safe interval then we simply do not know and bail out (returning
// false).
//
// Similarly we have to take into account the imprecision of 'w' when finding
// the closest representation of 'w'. If we have two potential
// representations, and one is closer to both w_low and w_high, then we know
// it is closer to the actual value v.
//
// By generating the digits of too_high we got the largest (closest to
// too_high) buffer that is still in the unsafe interval. In the case where
// w_high < buffer < too_high we try to decrement the buffer.
// This way the buffer approaches (rounds towards) w.
// There are 3 conditions that stop the decrementation process:
// 1) the buffer is already below w_high
// 2) decrementing the buffer would make it leave the unsafe interval
// 3) decrementing the buffer would yield a number below w_high and farther
// away than the current number. In other words:
// (buffer{-1} < w_high) && w_high - buffer{-1} > buffer - w_high
// Instead of using the buffer directly we use its distance to too_high.
// Conceptually rest ~= too_high - buffer
// We need to do the following tests in this order to avoid over- and
// underflows.
ASSERT(rest <= unsafe_interval);
while (rest < small_distance && // Negated condition 1
unsafe_interval - rest >= ten_kappa && // Negated condition 2
(rest + ten_kappa < small_distance || // buffer{-1} > w_high
small_distance - rest >= rest + ten_kappa - small_distance)) {
buffer[length - 1]--;
rest += ten_kappa;
}
// We have approached w+ as much as possible. We now test if approaching w-
// would require changing the buffer. If yes, then we have two possible
// representations close to w, but we cannot decide which one is closer.
if (rest < big_distance &&
unsafe_interval - rest >= ten_kappa &&
(rest + ten_kappa < big_distance ||
big_distance - rest > rest + ten_kappa - big_distance)) {
return false;
}
// Weeding test.
// The safe interval is [too_low + 2 ulp; too_high - 2 ulp]
// Since too_low = too_high - unsafe_interval this is equivalent to
// [too_high - unsafe_interval + 4 ulp; too_high - 2 ulp]
// Conceptually we have: rest ~= too_high - buffer
return (2 * unit <= rest) && (rest <= unsafe_interval - 4 * unit);
}
// Rounds the buffer upwards if the result is closer to v by possibly adding
// 1 to the buffer. If the precision of the calculation is not sufficient to
// round correctly, return false.
// The rounding might shift the whole buffer in which case the kappa is
// adjusted. For example "99", kappa = 3 might become "10", kappa = 4.
//
// If 2*rest > ten_kappa then the buffer needs to be round up.
// rest can have an error of +/- 1 unit. This function accounts for the
// imprecision and returns false, if the rounding direction cannot be
// unambiguously determined.
//
// Precondition: rest < ten_kappa.
static bool RoundWeedCounted(Vector<char> buffer,
int length,
uint64_t rest,
uint64_t ten_kappa,
uint64_t unit,
int* kappa) {
ASSERT(rest < ten_kappa);
// The following tests are done in a specific order to avoid overflows. They
// will work correctly with any uint64 values of rest < ten_kappa and unit.
//
// If the unit is too big, then we don't know which way to round. For example
// a unit of 50 means that the real number lies within rest +/- 50. If
// 10^kappa == 40 then there is no way to tell which way to round.
if (unit >= ten_kappa) return false;
// Even if unit is just half the size of 10^kappa we are already completely
// lost. (And after the previous test we know that the expression will not
// over/underflow.)
if (ten_kappa - unit <= unit) return false;
// If 2 * (rest + unit) <= 10^kappa we can safely round down.
if ((ten_kappa - rest > rest) && (ten_kappa - 2 * rest >= 2 * unit)) {
return true;
}
// If 2 * (rest - unit) >= 10^kappa, then we can safely round up.
if ((rest > unit) && (ten_kappa - (rest - unit) <= (rest - unit))) {
// Increment the last digit recursively until we find a non '9' digit.
buffer[length - 1]++;
for (int i = length - 1; i > 0; --i) {
if (buffer[i] != '0' + 10) break;
buffer[i] = '0';
buffer[i - 1]++;
}
// If the first digit is now '0'+ 10 we had a buffer with all '9's. With the
// exception of the first digit all digits are now '0'. Simply switch the
// first digit to '1' and adjust the kappa. Example: "99" becomes "10" and
// the power (the kappa) is increased.
if (buffer[0] == '0' + 10) {
buffer[0] = '1';
(*kappa) += 1;
}
return true;
}
return false;
}
// Returns the biggest power of ten that is less than or equal to the given
// number. We furthermore receive the maximum number of bits 'number' has.
//
// Returns power == 10^(exponent_plus_one-1) such that
// power <= number < power * 10.
// If number_bits == 0 then 0^(0-1) is returned.
// The number of bits must be <= 32.
// Precondition: number < (1 << (number_bits + 1)).
// Inspired by the method for finding an integer log base 10 from here:
// http://graphics.stanford.edu/~seander/bithacks.html#IntegerLog10
static unsigned int const kSmallPowersOfTen[] =
{0, 1, 10, 100, 1000, 10000, 100000, 1000000, 10000000, 100000000,
1000000000};
static void BiggestPowerTen(uint32_t number,
int number_bits,
uint32_t* power,
int* exponent_plus_one) {
ASSERT(number < (1u << (number_bits + 1)));
// 1233/4096 is approximately 1/lg(10).
int exponent_plus_one_guess = ((number_bits + 1) * 1233 >> 12);
// We increment to skip over the first entry in the kPowersOf10 table.
// Note: kPowersOf10[i] == 10^(i-1).
exponent_plus_one_guess++;
// We don't have any guarantees that 2^number_bits <= number.
if (number < kSmallPowersOfTen[exponent_plus_one_guess]) {
exponent_plus_one_guess--;
}
*power = kSmallPowersOfTen[exponent_plus_one_guess];
*exponent_plus_one = exponent_plus_one_guess;
}
// Generates the digits of input number w.
// w is a floating-point number (DiyFp), consisting of a significand and an
// exponent. Its exponent is bounded by kMinimalTargetExponent and
// kMaximalTargetExponent.
// Hence -60 <= w.e() <= -32.
//
// Returns false if it fails, in which case the generated digits in the buffer
// should not be used.
// Preconditions:
// * low, w and high are correct up to 1 ulp (unit in the last place). That
// is, their error must be less than a unit of their last digits.
// * low.e() == w.e() == high.e()
// * low < w < high, and taking into account their error: low~ <= high~
// * kMinimalTargetExponent <= w.e() <= kMaximalTargetExponent
// Postconditions: returns false if procedure fails.
// otherwise:
// * buffer is not null-terminated, but len contains the number of digits.
// * buffer contains the shortest possible decimal digit-sequence
// such that LOW < buffer * 10^kappa < HIGH, where LOW and HIGH are the
// correct values of low and high (without their error).
// * if more than one decimal representation gives the minimal number of
// decimal digits then the one closest to W (where W is the correct value
// of w) is chosen.
// Remark: this procedure takes into account the imprecision of its input
// numbers. If the precision is not enough to guarantee all the postconditions
// then false is returned. This usually happens rarely (~0.5%).
//
// Say, for the sake of example, that
// w.e() == -48, and w.f() == 0x1234567890abcdef
// w's value can be computed by w.f() * 2^w.e()
// We can obtain w's integral digits by simply shifting w.f() by -w.e().
// -> w's integral part is 0x1234
// w's fractional part is therefore 0x567890abcdef.
// Printing w's integral part is easy (simply print 0x1234 in decimal).
// In order to print its fraction we repeatedly multiply the fraction by 10 and
// get each digit. Example the first digit after the point would be computed by
// (0x567890abcdef * 10) >> 48. -> 3
// The whole thing becomes slightly more complicated because we want to stop
// once we have enough digits. That is, once the digits inside the buffer
// represent 'w' we can stop. Everything inside the interval low - high
// represents w. However we have to pay attention to low, high and w's
// imprecision.
static bool DigitGen(DiyFp low,
DiyFp w,
DiyFp high,
Vector<char> buffer,
int* length,
int* kappa) {
ASSERT(low.e() == w.e() && w.e() == high.e());
ASSERT(low.f() + 1 <= high.f() - 1);
ASSERT(kMinimalTargetExponent <= w.e() && w.e() <= kMaximalTargetExponent);
// low, w and high are imprecise, but by less than one ulp (unit in the last
// place).
// If we remove (resp. add) 1 ulp from low (resp. high) we are certain that
// the new numbers are outside of the interval we want the final
// representation to lie in.
// Inversely adding (resp. removing) 1 ulp from low (resp. high) would yield
// numbers that are certain to lie in the interval. We will use this fact
// later on.
// We will now start by generating the digits within the uncertain
// interval. Later we will weed out representations that lie outside the safe
// interval and thus _might_ lie outside the correct interval.
uint64_t unit = 1;
DiyFp too_low = DiyFp(low.f() - unit, low.e());
DiyFp too_high = DiyFp(high.f() + unit, high.e());
// too_low and too_high are guaranteed to lie outside the interval we want the
// generated number in.
DiyFp unsafe_interval = DiyFp::Minus(too_high, too_low);
// We now cut the input number into two parts: the integral digits and the
// fractionals. We will not write any decimal separator though, but adapt
// kappa instead.
// Reminder: we are currently computing the digits (stored inside the buffer)
// such that: too_low < buffer * 10^kappa < too_high
// We use too_high for the digit_generation and stop as soon as possible.
// If we stop early we effectively round down.
DiyFp one = DiyFp(static_cast<uint64_t>(1) << -w.e(), w.e());
// Division by one is a shift.
uint32_t integrals = static_cast<uint32_t>(too_high.f() >> -one.e());
// Modulo by one is an and.
uint64_t fractionals = too_high.f() & (one.f() - 1);
uint32_t divisor;
int divisor_exponent_plus_one;
BiggestPowerTen(integrals, DiyFp::kSignificandSize - (-one.e()),
&divisor, &divisor_exponent_plus_one);
*kappa = divisor_exponent_plus_one;
*length = 0;
// Loop invariant: buffer = too_high / 10^kappa (integer division)
// The invariant holds for the first iteration: kappa has been initialized
// with the divisor exponent + 1. And the divisor is the biggest power of ten
// that is smaller than integrals.
while (*kappa > 0) {
int digit = integrals / divisor;
ASSERT(digit <= 9);
buffer[*length] = static_cast<char>('0' + digit);
(*length)++;
integrals %= divisor;
(*kappa)--;
// Note that kappa now equals the exponent of the divisor and that the
// invariant thus holds again.
uint64_t rest =
(static_cast<uint64_t>(integrals) << -one.e()) + fractionals;
// Invariant: too_high = buffer * 10^kappa + DiyFp(rest, one.e())
// Reminder: unsafe_interval.e() == one.e()
if (rest < unsafe_interval.f()) {
// Rounding down (by not emitting the remaining digits) yields a number
// that lies within the unsafe interval.
return RoundWeed(buffer, *length, DiyFp::Minus(too_high, w).f(),
unsafe_interval.f(), rest,
static_cast<uint64_t>(divisor) << -one.e(), unit);
}
divisor /= 10;
}
// The integrals have been generated. We are at the point of the decimal
// separator. In the following loop we simply multiply the remaining digits by
// 10 and divide by one. We just need to pay attention to multiply associated
// data (like the interval or 'unit'), too.
// Note that the multiplication by 10 does not overflow, because w.e >= -60
// and thus one.e >= -60.
ASSERT(one.e() >= -60);
ASSERT(fractionals < one.f());
ASSERT(UINT64_2PART_C(0xFFFFFFFF, FFFFFFFF) / 10 >= one.f());
for (;;) {
fractionals *= 10;
unit *= 10;
unsafe_interval.set_f(unsafe_interval.f() * 10);
// Integer division by one.
int digit = static_cast<int>(fractionals >> -one.e());
ASSERT(digit <= 9);
buffer[*length] = static_cast<char>('0' + digit);
(*length)++;
fractionals &= one.f() - 1; // Modulo by one.
(*kappa)--;
if (fractionals < unsafe_interval.f()) {
return RoundWeed(buffer, *length, DiyFp::Minus(too_high, w).f() * unit,
unsafe_interval.f(), fractionals, one.f(), unit);
}
}
}
// Generates (at most) requested_digits digits of input number w.
// w is a floating-point number (DiyFp), consisting of a significand and an
// exponent. Its exponent is bounded by kMinimalTargetExponent and
// kMaximalTargetExponent.
// Hence -60 <= w.e() <= -32.
//
// Returns false if it fails, in which case the generated digits in the buffer
// should not be used.
// Preconditions:
// * w is correct up to 1 ulp (unit in the last place). That
// is, its error must be strictly less than a unit of its last digit.
// * kMinimalTargetExponent <= w.e() <= kMaximalTargetExponent
//
// Postconditions: returns false if procedure fails.
// otherwise:
// * buffer is not null-terminated, but length contains the number of
// digits.
// * the representation in buffer is the most precise representation of
// requested_digits digits.
// * buffer contains at most requested_digits digits of w. If there are less
// than requested_digits digits then some trailing '0's have been removed.
// * kappa is such that
// w = buffer * 10^kappa + eps with |eps| < 10^kappa / 2.
//
// Remark: This procedure takes into account the imprecision of its input
// numbers. If the precision is not enough to guarantee all the postconditions
// then false is returned. This usually happens rarely, but the failure-rate
// increases with higher requested_digits.
static bool DigitGenCounted(DiyFp w,
int requested_digits,
Vector<char> buffer,
int* length,
int* kappa) {
ASSERT(kMinimalTargetExponent <= w.e() && w.e() <= kMaximalTargetExponent);
ASSERT(kMinimalTargetExponent >= -60);
ASSERT(kMaximalTargetExponent <= -32);
// w is assumed to have an error less than 1 unit. Whenever w is scaled we
// also scale its error.
uint64_t w_error = 1;
// We cut the input number into two parts: the integral digits and the
// fractional digits. We don't emit any decimal separator, but adapt kappa
// instead. Example: instead of writing "1.2" we put "12" into the buffer and
// increase kappa by 1.
DiyFp one = DiyFp(static_cast<uint64_t>(1) << -w.e(), w.e());
// Division by one is a shift.
uint32_t integrals = static_cast<uint32_t>(w.f() >> -one.e());
// Modulo by one is an and.
uint64_t fractionals = w.f() & (one.f() - 1);
uint32_t divisor;
int divisor_exponent_plus_one;
BiggestPowerTen(integrals, DiyFp::kSignificandSize - (-one.e()),
&divisor, &divisor_exponent_plus_one);
*kappa = divisor_exponent_plus_one;
*length = 0;
// Loop invariant: buffer = w / 10^kappa (integer division)
// The invariant holds for the first iteration: kappa has been initialized
// with the divisor exponent + 1. And the divisor is the biggest power of ten
// that is smaller than 'integrals'.
while (*kappa > 0) {
int digit = integrals / divisor;
ASSERT(digit <= 9);
buffer[*length] = static_cast<char>('0' + digit);
(*length)++;
requested_digits--;
integrals %= divisor;
(*kappa)--;
// Note that kappa now equals the exponent of the divisor and that the
// invariant thus holds again.
if (requested_digits == 0) break;
divisor /= 10;
}
if (requested_digits == 0) {
uint64_t rest =
(static_cast<uint64_t>(integrals) << -one.e()) + fractionals;
return RoundWeedCounted(buffer, *length, rest,
static_cast<uint64_t>(divisor) << -one.e(), w_error,
kappa);
}
// The integrals have been generated. We are at the point of the decimal
// separator. In the following loop we simply multiply the remaining digits by
// 10 and divide by one. We just need to pay attention to multiply associated
// data (the 'unit'), too.
// Note that the multiplication by 10 does not overflow, because w.e >= -60
// and thus one.e >= -60.
ASSERT(one.e() >= -60);
ASSERT(fractionals < one.f());
ASSERT(UINT64_2PART_C(0xFFFFFFFF, FFFFFFFF) / 10 >= one.f());
while (requested_digits > 0 && fractionals > w_error) {
fractionals *= 10;
w_error *= 10;
// Integer division by one.
int digit = static_cast<int>(fractionals >> -one.e());
ASSERT(digit <= 9);
buffer[*length] = static_cast<char>('0' + digit);
(*length)++;
requested_digits--;
fractionals &= one.f() - 1; // Modulo by one.
(*kappa)--;
}
if (requested_digits != 0) return false;
return RoundWeedCounted(buffer, *length, fractionals, one.f(), w_error,
kappa);
}
// Provides a decimal representation of v.
// Returns true if it succeeds, otherwise the result cannot be trusted.
// There will be *length digits inside the buffer (not null-terminated).
// If the function returns true then
// v == (double) (buffer * 10^decimal_exponent).
// The digits in the buffer are the shortest representation possible: no
// 0.09999999999999999 instead of 0.1. The shorter representation will even be
// chosen even if the longer one would be closer to v.
// The last digit will be closest to the actual v. That is, even if several
// digits might correctly yield 'v' when read again, the closest will be
// computed.
static bool Grisu3(double v,
FastDtoaMode mode,
Vector<char> buffer,
int* length,
int* decimal_exponent) {
DiyFp w = Double(v).AsNormalizedDiyFp();
// boundary_minus and boundary_plus are the boundaries between v and its
// closest floating-point neighbors. Any number strictly between
// boundary_minus and boundary_plus will round to v when convert to a double.
// Grisu3 will never output representations that lie exactly on a boundary.
DiyFp boundary_minus, boundary_plus;
if (mode == FAST_DTOA_SHORTEST) {
Double(v).NormalizedBoundaries(&boundary_minus, &boundary_plus);
} else {
ASSERT(mode == FAST_DTOA_SHORTEST_SINGLE);
float single_v = static_cast<float>(v);
Single(single_v).NormalizedBoundaries(&boundary_minus, &boundary_plus);
}
ASSERT(boundary_plus.e() == w.e());
DiyFp ten_mk; // Cached power of ten: 10^-k
int mk; // -k
int ten_mk_minimal_binary_exponent =
kMinimalTargetExponent - (w.e() + DiyFp::kSignificandSize);
int ten_mk_maximal_binary_exponent =
kMaximalTargetExponent - (w.e() + DiyFp::kSignificandSize);
PowersOfTenCache::GetCachedPowerForBinaryExponentRange(
ten_mk_minimal_binary_exponent,
ten_mk_maximal_binary_exponent,
&ten_mk, &mk);
ASSERT((kMinimalTargetExponent <= w.e() + ten_mk.e() +
DiyFp::kSignificandSize) &&
(kMaximalTargetExponent >= w.e() + ten_mk.e() +
DiyFp::kSignificandSize));
// Note that ten_mk is only an approximation of 10^-k. A DiyFp only contains a
// 64 bit significand and ten_mk is thus only precise up to 64 bits.
// The DiyFp::Times procedure rounds its result, and ten_mk is approximated
// too. The variable scaled_w (as well as scaled_boundary_minus/plus) are now
// off by a small amount.
// In fact: scaled_w - w*10^k < 1ulp (unit in the last place) of scaled_w.
// In other words: let f = scaled_w.f() and e = scaled_w.e(), then
// (f-1) * 2^e < w*10^k < (f+1) * 2^e
DiyFp scaled_w = DiyFp::Times(w, ten_mk);
ASSERT(scaled_w.e() ==
boundary_plus.e() + ten_mk.e() + DiyFp::kSignificandSize);
// In theory it would be possible to avoid some recomputations by computing
// the difference between w and boundary_minus/plus (a power of 2) and to
// compute scaled_boundary_minus/plus by subtracting/adding from
// scaled_w. However the code becomes much less readable and the speed
// enhancements are not terriffic.
DiyFp scaled_boundary_minus = DiyFp::Times(boundary_minus, ten_mk);
DiyFp scaled_boundary_plus = DiyFp::Times(boundary_plus, ten_mk);
// DigitGen will generate the digits of scaled_w. Therefore we have
// v == (double) (scaled_w * 10^-mk).
// Set decimal_exponent == -mk and pass it to DigitGen. If scaled_w is not an
// integer than it will be updated. For instance if scaled_w == 1.23 then
// the buffer will be filled with "123" und the decimal_exponent will be
// decreased by 2.
int kappa;
bool result = DigitGen(scaled_boundary_minus, scaled_w, scaled_boundary_plus,
buffer, length, &kappa);
*decimal_exponent = -mk + kappa;
return result;
}
// The "counted" version of grisu3 (see above) only generates requested_digits
// number of digits. This version does not generate the shortest representation,
// and with enough requested digits 0.1 will at some point print as 0.9999999...
// Grisu3 is too imprecise for real halfway cases (1.5 will not work) and
// therefore the rounding strategy for halfway cases is irrelevant.
static bool Grisu3Counted(double v,
int requested_digits,
Vector<char> buffer,
int* length,
int* decimal_exponent) {
DiyFp w = Double(v).AsNormalizedDiyFp();
DiyFp ten_mk; // Cached power of ten: 10^-k
int mk; // -k
int ten_mk_minimal_binary_exponent =
kMinimalTargetExponent - (w.e() + DiyFp::kSignificandSize);
int ten_mk_maximal_binary_exponent =
kMaximalTargetExponent - (w.e() + DiyFp::kSignificandSize);
PowersOfTenCache::GetCachedPowerForBinaryExponentRange(
ten_mk_minimal_binary_exponent,
ten_mk_maximal_binary_exponent,
&ten_mk, &mk);
ASSERT((kMinimalTargetExponent <= w.e() + ten_mk.e() +
DiyFp::kSignificandSize) &&
(kMaximalTargetExponent >= w.e() + ten_mk.e() +
DiyFp::kSignificandSize));
// Note that ten_mk is only an approximation of 10^-k. A DiyFp only contains a
// 64 bit significand and ten_mk is thus only precise up to 64 bits.
// The DiyFp::Times procedure rounds its result, and ten_mk is approximated
// too. The variable scaled_w (as well as scaled_boundary_minus/plus) are now
// off by a small amount.
// In fact: scaled_w - w*10^k < 1ulp (unit in the last place) of scaled_w.
// In other words: let f = scaled_w.f() and e = scaled_w.e(), then
// (f-1) * 2^e < w*10^k < (f+1) * 2^e
DiyFp scaled_w = DiyFp::Times(w, ten_mk);
// We now have (double) (scaled_w * 10^-mk).
// DigitGen will generate the first requested_digits digits of scaled_w and
// return together with a kappa such that scaled_w ~= buffer * 10^kappa. (It
// will not always be exactly the same since DigitGenCounted only produces a
// limited number of digits.)
int kappa;
bool result = DigitGenCounted(scaled_w, requested_digits,
buffer, length, &kappa);
*decimal_exponent = -mk + kappa;
return result;
}
bool FastDtoa(double v,
FastDtoaMode mode,
int requested_digits,
Vector<char> buffer,
int* length,
int* decimal_point) {
ASSERT(v > 0);
ASSERT(!Double(v).IsSpecial());
bool result = false;
int decimal_exponent = 0;
switch (mode) {
case FAST_DTOA_SHORTEST:
case FAST_DTOA_SHORTEST_SINGLE:
result = Grisu3(v, mode, buffer, length, &decimal_exponent);
break;
case FAST_DTOA_PRECISION:
result = Grisu3Counted(v, requested_digits,
buffer, length, &decimal_exponent);
break;
default:
UNREACHABLE();
}
if (result) {
*decimal_point = *length + decimal_exponent;
buffer[*length] = '\0';
}
return result;
}
} // namespace double_conversion