Log in
Register
Menu
Log in
Register
Home
What's new
Latest activity
Authors
Forums
New posts
Search forums
What's new
New posts
Latest activity
Members
Current visitors
New posts
Search forums
Menu
Log in
Register
Install the app
Install
Forums
Satellite TV receivers & systems support forums
Satellite PC Card Receivers, Internet by Satellite
neumoDVB
JavaScript is disabled. For a better experience, please enable JavaScript in your browser before proceeding.
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an
alternative browser
.
Reply to thread
Message
<blockquote data-quote="deeptho" data-source="post: 1154839" data-attributes="member: 215446"><p>Here is a list of the changes in the next release - Part 1</p><h2>Changes in version neumodvb-1.4</h2><h3>Main changes</h3> <ul> <li data-xf-list-type="ul">Allow scanning all existing muxes on multiple satellites, while selecting muxes based on polarisation and band.</li> <li data-xf-list-type="ul">Allow scanning spectral bands on multiple satellites, while selecting bands based on polarisation and band and on frequency range.</li> <li data-xf-list-type="ul">Have muxinfo display more info during scanning.</li> <li data-xf-list-type="ul">Satellite list now shows information on last band scan in a multi-row format.</li> <li data-xf-list-type="ul">Improved code for handling conflicts during subscriptions.</li> <li data-xf-list-type="ul">Allow mux and service subscription to share same active mux.</li> <li data-xf-list-type="ul">Various bug fixes related to incorrect selection of LNBs or not powering up LNBs or selecting bands when multples muxes are tuned in parallel. This also leads to more reliable scanning.</li> <li data-xf-list-type="ul">Refactor scan code to make it less complicated and easier to maintain.</li> <li data-xf-list-type="ul">Fixes for recovering recordings when neumoDVB is restarted while a recording is in progress.</li> <li data-xf-list-type="ul">Add autorec code, which checks all new EPG data and creates recording records based on matching program name, story, start time, end time…</li> <li data-xf-list-type="ul">sat_list is now organized per frequency band dvbc and dvbt are no longer included in that list. Having a different satellite entry per frequency band makes it easier to display only those satellites supporting a specific band.</li> <li data-xf-list-type="ul">Use ch_order=65535 for new services so that they appear at end of service list and display 65535 as empty string. The end result is that for services without a ch_order are now displayed with a blank channel number and at the end of the list, as they are often not of interest to the user.</li> <li data-xf-list-type="ul">Distinguish between cur_lnb_pos (what position does lnb point to) and cur_sat_pos (what as the last network used on lnb). The difference is important when multiple closely spaced satellites are received by the same lnb.</li> <li data-xf-list-type="ul">Channel epg list has been improved: it allows filtering for a specifuic service. The info window has been removed. Instead, the story is shown using multiple lines in a separate column.</li> <li data-xf-list-type="ul">Much faster Sky UK and Freesat EPG processing due to various improvements to internal huffman decoders. All Sky UK epg can now be grabbed in less than 30 seconds.</li> <li data-xf-list-type="ul">The EPG code, and some other code now also runs in parallel when scanning multiple muxes at once. The lmdb database still leads to some serialization.</li> <li data-xf-list-type="ul">All debugging code now uses libfmt, leading to faster logging and more informative log messages with cleaner code.</li> <li data-xf-list-type="ul">Internal database code has been improved to allow parallel calling from multiple threads without unneeded committing of transactions.</li> <li data-xf-list-type="ul">Muxes now use the new mux_id as part of new primary for mux. This makes it easier to avoid bugs causing the wrong mux record to be updated or deleted, when multiple muxes overlap.</li> <li data-xf-list-type="ul">In the service list the string column mux_desc has now been replaced with separate frequency and polarisation columns, allowing filtering and sorting by frequency or polarisation.</li> <li data-xf-list-type="ul">The database now also stores schema_version in database records and in data structures, to be able to detect major schema upgrades.</li> </ul><h3>Spectrum scan and positioner dialog</h3> <ul> <li data-xf-list-type="ul">Send lnb error to tune_mux_panel rather than handle it via global subscriber.</li> <li data-xf-list-type="ul">Improved error message when lnb action fails.</li> <li data-xf-list-type="ul">When tune_mux_panel action fails, handle resulting error in spectrum_dialog or positioner_dialog.</li> <li data-xf-list-type="ul">Reset get spectrum button if spectrum scan fails.</li> <li data-xf-list-type="ul">Bug: spectra acquired within one minute not treated as the same in GUI, resulting in hiding of first spectrum when second is displayed.</li> <li data-xf-list-type="ul">Move start_time from spectrum_options to internal fe state.</li> <li data-xf-list-type="ul">Make get_spectrum also return peak and freq data to scan notification code.</li> <li data-xf-list-type="ul">Bug: Indexing of spectra no longer correct.</li> </ul><h3>More reliable tuning and preventing conflicts between parallel tunes</h3> <ul> <li data-xf-list-type="ul">Incorrect comparison when no service was reserved.</li> <li data-xf-list-type="ul">New fe_subscribe code.</li> <li data-xf-list-type="ul">Report error message when lnb cannot be subscribed exclusively.</li> <li data-xf-list-type="ul">Bug: mux list popup in positioner dialog shows Ka band muxes when LNB is of Ku type.</li> <li data-xf-list-type="ul">Bug: frontend info not retrieved for adapters in use and file descriptor not closed.</li> <li data-xf-list-type="ul">Bug: Incorrect live_service processing causes deletion of live buffers in use.</li> <li data-xf-list-type="ul">Faster channel changing: instead of first calling unsubscribe from neumompv and then resubscribing, let tuner code do the unsubscribing. This can often reuse an existing active adapter and avoids closing and then reopening a frontend.</li> <li data-xf-list-type="ul">Bug: active_adapter released when unsubscribing one of multiple subscriptions.</li> <li data-xf-list-type="ul">Bug: cannot call start_running in constructor of active_adapter.</li> <li data-xf-list-type="ul">Identify live_service_t by owner and subscription_id; remove epg field in live_service_t to simplify code.</li> <li data-xf-list-type="ul">Incorrect handling of lock loss.</li> <li data-xf-list-type="ul">Confusion between ret and errno.</li> <li data-xf-list-type="ul">Bug: incorrect handling of resource conflicts.</li> <li data-xf-list-type="ul">Bug: incorrect handling of unsubscription when multiple subscriptions use same fe.</li> <li data-xf-list-type="ul">Bug: when subscription fails, active_adapter and subscription are not released.</li> <li data-xf-list-type="ul">Bug: incorrect sending diseqc messages and changing voltage/pol during retune.</li> <li data-xf-list-type="ul">Bug: active_adapter recreated even when adapter remains the same. New fe_subscribe code.</li> <li data-xf-list-type="ul">Bug: assertion when tuning dvbt mux.</li> <li data-xf-list-type="ul">Increase sleep times after changing voltage or rf input.</li> <li data-xf-list-type="ul">Bug: when switching to a different lnb on a frontend, we need to allow for sufficient power up time to prevent tuning from failing.</li> <li data-xf-list-type="ul">Do not power down tuner voltage before switching to a new lnb. Instead start with the old voltage (faster and better).</li> <li data-xf-list-type="ul">Bug: diseqc commands not sent when switching to the same lnb via another card or rf_input (e.g., to scan another polarisation).</li> <li data-xf-list-type="ul">Also send voltage commands afer rf_input has changed; otherwise tuning may fail.</li> <li data-xf-list-type="ul">Bug: not checking if subscription uses same sat/pol/band when checking if frontend in use can be used leads to reusing subscription from another lnb.</li> <li data-xf-list-type="ul">Bug: when reusing the same adapter, tuning uses old lnb.</li> <li data-xf-list-type="ul">Bug: incorrect decision that card cannot be used.</li> <li data-xf-list-type="ul">Bug: incorrect detection of need_diseqc.</li> <li data-xf-list-type="ul">Bug: sat_pos not taken into account when checking if subscription can be reused.</li> <li data-xf-list-type="ul">Bug: subscribe not handling properly the case where a subscription subscribes to a service (or to the mux itself) on the same mux as its old subscription.</li> <li data-xf-list-type="ul">Unify subscribe_mux and matching_existing_subscription code for all dvb types.</li> <li data-xf-list-type="ul">Bug: reserve_fe_in_use not storing mux_key in subscription when subscribing to mux (instead of service).</li> <li data-xf-list-type="ul">Simplify subscriber notifications.</li> <li data-xf-list-type="ul">Allow non exclusive satellite band reservations.</li> <li data-xf-list-type="ul">drop may_control_lnb, may_move_dish, need_blind_tune, need_spectrum and loc a function arguments; use tune_options instead. Also remove them from subscription_ret_t. Remove tune_pars_t and replace with tune_optons_t.</li> <li data-xf-list-type="ul">move several function parameters into tune_options_t; specifically include usals_location, dish_move_penalty and resource_reuse_bonus in tune_options.</li> <li data-xf-list-type="ul">Switch most “subscribe” calls to using tune_options_t parameter.</li> <li data-xf-list-type="ul">Bug: deadlock when reading last_signal_info.</li> <li data-xf-list-type="ul">Incorrect debug message when overriding obviously incorrect si data with driver data.</li> <li data-xf-list-type="ul">Simplify update_mux and its callbacks.</li> <li data-xf-list-type="ul">constness of merge_muxes argument.</li> </ul><h3>Improved mux and satellite band scanning</h3> <ul> <li data-xf-list-type="ul">Add allowed_rf_paths, allowed_dishes, and allowed_card_mac_addresses to tune_options to more finely control what resources can be used in tuning.</li> <li data-xf-list-type="ul">Last blind scan setting in positioner and spectrum dialog now remembered as future default.</li> <li data-xf-list-type="ul">Bug: initial scan_report not shown.</li> <li data-xf-list-type="ul">Improved decisions on whether to use mux data from database while scanning spectral peaks.</li> <li data-xf-list-type="ul">Allow add_spectral_peaks to be also called with data from fe_monitor and handle the way that received spectra asre processed: send them also to scanner_t code.</li> <li data-xf-list-type="ul">Bug: incorrect scan statistics.</li> <li data-xf-list-type="ul">Bug: ensure that canceling subscription correctly ends scan.</li> <li data-xf-list-type="ul">Bug: mux_common and scan_id lost when rescanning peak.</li> <li data-xf-list-type="ul">Bug: incorrect rescanning on scan_result_t:<img src="/styles/default/xenforo/smilies/emojione/thumbsdown.png" class="smilie" loading="lazy" alt=":N" title="Thumbs Down :N" data-shortname=":N" />OTS.</li> <li data-xf-list-type="ul">Gather scan_stats at point of change directly instead of returning them from function calls.</li> <li data-xf-list-type="ul">Bug: incorrect detection of subscribing to a frequency peak.</li> <li data-xf-list-type="ul">No longer include scan_stats in scan_report. Rename scan_report to scan mux_end_report_t. Ensure that notify_scan_mux_end is also called for muxes that fail to tune at subscription time. Refactor scan_loop.</li> <li data-xf-list-type="ul">Separate code for scanning peaks and muxes.</li> <li data-xf-list-type="ul">Make all notification functions thread safe, while still notifying python layers without delay whenever possible. Harmonize naming of notification functions.</li> <li data-xf-list-type="ul">Safe deletion in ~scanner_t.</li> <li data-xf-list-type="ul">Bug: when unsubscribing, data stored in fe_t and in database becomes inconsistent, leading to never ending scan.</li> <li data-xf-list-type="ul">Bug: incorrect handling of incomplete BAT tables.</li> <li data-xf-list-type="ul">Ignore DefaultAuthorityDescriptorTag to remove many log messages.</li> <li data-xf-list-type="ul">Bug: incorrect frequency from si data overwrites correct data from driver.</li> <li data-xf-list-type="ul">Bug: when reusing active_adapter old fds remain open, leading to incorrect early notification of scanner.</li> <li data-xf-list-type="ul">Bug: last_signal_info and setting lock status in frontend.cc was performed in two steps, leading tuner_thread to see inconsistent information (is_dvb not yet set when lock is reported).</li> <li data-xf-list-type="ul">Bug: incorrect reuse of existing subscription when scanning spectral peaks.</li> <li data-xf-list-type="ul">Ensure generated mux_id always > 0.</li> <li data-xf-list-type="ul">Bug: incorrect growing of vectors of plain old data. Leads to bouquet processing incorrectly expiring entries on 28.2E.</li> <li data-xf-list-type="ul">Bug: when reusing active_adapter old fds remain open, leading to incorrect early notification of scanner.</li> <li data-xf-list-type="ul">Ignore matype for dvbc and dvbt</li> </ul></blockquote><p></p>
[QUOTE="deeptho, post: 1154839, member: 215446"] Here is a list of the changes in the next release - Part 1 [HEADING=1]Changes in version neumodvb-1.4[/HEADING] [HEADING=2]Main changes[/HEADING] [LIST] [*]Allow scanning all existing muxes on multiple satellites, while selecting muxes based on polarisation and band. [*]Allow scanning spectral bands on multiple satellites, while selecting bands based on polarisation and band and on frequency range. [*]Have muxinfo display more info during scanning. [*]Satellite list now shows information on last band scan in a multi-row format. [*]Improved code for handling conflicts during subscriptions. [*]Allow mux and service subscription to share same active mux. [*]Various bug fixes related to incorrect selection of LNBs or not powering up LNBs or selecting bands when multples muxes are tuned in parallel. This also leads to more reliable scanning. [*]Refactor scan code to make it less complicated and easier to maintain. [*]Fixes for recovering recordings when neumoDVB is restarted while a recording is in progress. [*]Add autorec code, which checks all new EPG data and creates recording records based on matching program name, story, start time, end time… [*]sat_list is now organized per frequency band dvbc and dvbt are no longer included in that list. Having a different satellite entry per frequency band makes it easier to display only those satellites supporting a specific band. [*]Use ch_order=65535 for new services so that they appear at end of service list and display 65535 as empty string. The end result is that for services without a ch_order are now displayed with a blank channel number and at the end of the list, as they are often not of interest to the user. [*]Distinguish between cur_lnb_pos (what position does lnb point to) and cur_sat_pos (what as the last network used on lnb). The difference is important when multiple closely spaced satellites are received by the same lnb. [*]Channel epg list has been improved: it allows filtering for a specifuic service. The info window has been removed. Instead, the story is shown using multiple lines in a separate column. [*]Much faster Sky UK and Freesat EPG processing due to various improvements to internal huffman decoders. All Sky UK epg can now be grabbed in less than 30 seconds. [*]The EPG code, and some other code now also runs in parallel when scanning multiple muxes at once. The lmdb database still leads to some serialization. [*]All debugging code now uses libfmt, leading to faster logging and more informative log messages with cleaner code. [*]Internal database code has been improved to allow parallel calling from multiple threads without unneeded committing of transactions. [*]Muxes now use the new mux_id as part of new primary for mux. This makes it easier to avoid bugs causing the wrong mux record to be updated or deleted, when multiple muxes overlap. [*]In the service list the string column mux_desc has now been replaced with separate frequency and polarisation columns, allowing filtering and sorting by frequency or polarisation. [*]The database now also stores schema_version in database records and in data structures, to be able to detect major schema upgrades. [/LIST] [HEADING=2]Spectrum scan and positioner dialog[/HEADING] [LIST] [*]Send lnb error to tune_mux_panel rather than handle it via global subscriber. [*]Improved error message when lnb action fails. [*]When tune_mux_panel action fails, handle resulting error in spectrum_dialog or positioner_dialog. [*]Reset get spectrum button if spectrum scan fails. [*]Bug: spectra acquired within one minute not treated as the same in GUI, resulting in hiding of first spectrum when second is displayed. [*]Move start_time from spectrum_options to internal fe state. [*]Make get_spectrum also return peak and freq data to scan notification code. [*]Bug: Indexing of spectra no longer correct. [/LIST] [HEADING=2]More reliable tuning and preventing conflicts between parallel tunes[/HEADING] [LIST] [*]Incorrect comparison when no service was reserved. [*]New fe_subscribe code. [*]Report error message when lnb cannot be subscribed exclusively. [*]Bug: mux list popup in positioner dialog shows Ka band muxes when LNB is of Ku type. [*]Bug: frontend info not retrieved for adapters in use and file descriptor not closed. [*]Bug: Incorrect live_service processing causes deletion of live buffers in use. [*]Faster channel changing: instead of first calling unsubscribe from neumompv and then resubscribing, let tuner code do the unsubscribing. This can often reuse an existing active adapter and avoids closing and then reopening a frontend. [*]Bug: active_adapter released when unsubscribing one of multiple subscriptions. [*]Bug: cannot call start_running in constructor of active_adapter. [*]Identify live_service_t by owner and subscription_id; remove epg field in live_service_t to simplify code. [*]Incorrect handling of lock loss. [*]Confusion between ret and errno. [*]Bug: incorrect handling of resource conflicts. [*]Bug: incorrect handling of unsubscription when multiple subscriptions use same fe. [*]Bug: when subscription fails, active_adapter and subscription are not released. [*]Bug: incorrect sending diseqc messages and changing voltage/pol during retune. [*]Bug: active_adapter recreated even when adapter remains the same. New fe_subscribe code. [*]Bug: assertion when tuning dvbt mux. [*]Increase sleep times after changing voltage or rf input. [*]Bug: when switching to a different lnb on a frontend, we need to allow for sufficient power up time to prevent tuning from failing. [*]Do not power down tuner voltage before switching to a new lnb. Instead start with the old voltage (faster and better). [*]Bug: diseqc commands not sent when switching to the same lnb via another card or rf_input (e.g., to scan another polarisation). [*]Also send voltage commands afer rf_input has changed; otherwise tuning may fail. [*]Bug: not checking if subscription uses same sat/pol/band when checking if frontend in use can be used leads to reusing subscription from another lnb. [*]Bug: when reusing the same adapter, tuning uses old lnb. [*]Bug: incorrect decision that card cannot be used. [*]Bug: incorrect detection of need_diseqc. [*]Bug: sat_pos not taken into account when checking if subscription can be reused. [*]Bug: subscribe not handling properly the case where a subscription subscribes to a service (or to the mux itself) on the same mux as its old subscription. [*]Unify subscribe_mux and matching_existing_subscription code for all dvb types. [*]Bug: reserve_fe_in_use not storing mux_key in subscription when subscribing to mux (instead of service). [*]Simplify subscriber notifications. [*]Allow non exclusive satellite band reservations. [*]drop may_control_lnb, may_move_dish, need_blind_tune, need_spectrum and loc a function arguments; use tune_options instead. Also remove them from subscription_ret_t. Remove tune_pars_t and replace with tune_optons_t. [*]move several function parameters into tune_options_t; specifically include usals_location, dish_move_penalty and resource_reuse_bonus in tune_options. [*]Switch most “subscribe” calls to using tune_options_t parameter. [*]Bug: deadlock when reading last_signal_info. [*]Incorrect debug message when overriding obviously incorrect si data with driver data. [*]Simplify update_mux and its callbacks. [*]constness of merge_muxes argument. [/LIST] [HEADING=2]Improved mux and satellite band scanning[/HEADING] [LIST] [*]Add allowed_rf_paths, allowed_dishes, and allowed_card_mac_addresses to tune_options to more finely control what resources can be used in tuning. [*]Last blind scan setting in positioner and spectrum dialog now remembered as future default. [*]Bug: initial scan_report not shown. [*]Improved decisions on whether to use mux data from database while scanning spectral peaks. [*]Allow add_spectral_peaks to be also called with data from fe_monitor and handle the way that received spectra asre processed: send them also to scanner_t code. [*]Bug: incorrect scan statistics. [*]Bug: ensure that canceling subscription correctly ends scan. [*]Bug: mux_common and scan_id lost when rescanning peak. [*]Bug: incorrect rescanning on scan_result_t::NOTS. [*]Gather scan_stats at point of change directly instead of returning them from function calls. [*]Bug: incorrect detection of subscribing to a frequency peak. [*]No longer include scan_stats in scan_report. Rename scan_report to scan mux_end_report_t. Ensure that notify_scan_mux_end is also called for muxes that fail to tune at subscription time. Refactor scan_loop. [*]Separate code for scanning peaks and muxes. [*]Make all notification functions thread safe, while still notifying python layers without delay whenever possible. Harmonize naming of notification functions. [*]Safe deletion in ~scanner_t. [*]Bug: when unsubscribing, data stored in fe_t and in database becomes inconsistent, leading to never ending scan. [*]Bug: incorrect handling of incomplete BAT tables. [*]Ignore DefaultAuthorityDescriptorTag to remove many log messages. [*]Bug: incorrect frequency from si data overwrites correct data from driver. [*]Bug: when reusing active_adapter old fds remain open, leading to incorrect early notification of scanner. [*]Bug: last_signal_info and setting lock status in frontend.cc was performed in two steps, leading tuner_thread to see inconsistent information (is_dvb not yet set when lock is reported). [*]Bug: incorrect reuse of existing subscription when scanning spectral peaks. [*]Ensure generated mux_id always > 0. [*]Bug: incorrect growing of vectors of plain old data. Leads to bouquet processing incorrectly expiring entries on 28.2E. [*]Bug: when reusing active_adapter old fds remain open, leading to incorrect early notification of scanner. [*]Ignore matype for dvbc and dvbt [/LIST] [/QUOTE]
Insert quotes…
Verification
Post reply
Forums
Satellite TV receivers & systems support forums
Satellite PC Card Receivers, Internet by Satellite
neumoDVB
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
Accept
Learn more…
Top
