[Pkg-cli-apps-commits] [SCM] openbve branch, master, updated. debian/1.4.0.0-1-29-gdec261f
Paul Sladen
debian at paul.sladen.org
Wed Jan 18 12:14:22 UTC 2012
The following commit has been merged in the master branch:
commit c650ca51cb84744243ab4240d70c1ce87578fd22
Author: Paul Sladen <debian at paul.sladen.org>
Date: Wed Jan 18 12:28:49 2012 +0100
Merge 'known-issues.txt' into main 'README.Debian'
diff --git a/debian/README.Debian b/debian/README.Debian
index d9ab3f0..4032723 100644
--- a/debian/README.Debian
+++ b/debian/README.Debian
@@ -22,3 +22,171 @@ located under /usr/share/games/bve/ --- or vice-versa.
Have fun, and try not to crash!
+
+Known bugs/issues with running OpenBVE on Linux/Ubuntu/Debian.
+==============================================================
+
+This list was originally held in 'known-issues.txt'.
+
+Fixed and merged by upstream:
+-----------------------------
+2. If you have texture flashing issues; this is because the C-based
+OpenGL code was receiving pointers to single integers from the
+OpenBVE. Fixed by passing a array. Fixed upstream by passing
+components separately.
+
+3. Crashes out whilst trying to promisicously read non-accessible
+directories (eg. '/lost+found' or '~/.dbus'). It is hunting
+for directories containing a valid 'train.dat' description file.
+Kludged for the moment with try{}catch{} to not die.
+
+15. Window manager icon loading was not handled.
+
+5. Non-existant audio; sound should work the first time, but after
+quitting OpenAL complains about uncleaned-up buffers. You will
+likely need to 'kill -9 pulseaudio', although this may not recover
+the situation. Something in OpenAL/SDL isn't being cleanly shutdown.
+You may get the occasional sound afterwards on second/third attempts
+but not the full multi-channel surround experience.
+
+6. Captialisation. The routes have a combination of UPPER, lower, and
+MiXeD cAsE filenames which is a problem on Unix. The file-loading
+routinues need modifying to search for alterntively cased files if
+the filename as spelt can't be found.
+
+7. The London Undergrond cab 'LT_C69_77' appears in the centre of
+the screen in OpenBVE, meaning that you can't see where you're going.
+As a work-around, you may able to unlock the cam and move yourself up
+a bit to look over the top of the console. Sort-of-fixed in 0.9.5.5.
+(Though you will need to unlock the camera and zoom out to see any of
+the panel). This is to do with the camera-restriction box algorithm
+used in OpenBVE.
+
+8. The sprites Gauges in cabs displayed are not displayed correctly,
+instead showing a different/stretched part of the overlay with blue
+tinting from the edges of the alpha mask. Clipping/position broken.
+You will have to rely on the small rectangle OpenBVE overlays
+('EMG/B3/B2/B1/N/P1/P2/P3') instead.
+
+21. 'Security system warning', this refers to signalling system, not
+to the program trying to perform anything naughty! As extension DLLs
+are not being available on Linux. Patched to only display a warning
+*if* an 'ats.cfg' was present . This avoids the Issues dialogue on an
+otherwise error-free route/train.
+
+22. Hidden dot files/folders (those starting with '.') are clutter
+the filemanager. Patched to skip listing dot-files; you can
+still type the full name in. May be worth restricting to only
+!Platform.Windows.
+
+26. British English and flag should be en-GB/GB (not 'en-UK/UK').
+
+21. In full-screen on Linux, the mouse-cursor (pointer) remains visible in
+the centre of the screen.
+
+18. ExitToMainMenu causes a fork() and return (leaving the user at the
+command prompt and daemonising the second instance into the background)
+rather than using exec() or restarting itself. This crashes OpenBVE.
+It needs openbve to pass "/fhs" back to itself if it was handed it.
+
+4. Trying to write the directory the executable is in; On Unix this is
+likely to be owned by route. Modified to use the CLR SpecialFolder
+functions and save configuration data in the right places.
+
+16. After selecting an exterior camera view and then switching back to
+the interior camera; the cab/panel overlay appears in a different position.
+
+12. Language Locale not automatically set by default; patched to use
+System.Globalization.CultureInfo.CurrentCulture.Name as default, but
+needs refining to cope with 'll_CC' not just 'll' language codes.
+This is done on first-load but then forcifully saved in the options
+even if the user has not manually selected a language (prevents
+future automatic language selection working).
+
+20. No way reset the Train (and camera) to be upright after a crash. You
+should drive more carefully next time!
+
+29. When running on non-Win32 systems where Win32 plugin loading is not
+attempted, the message given may be the current one. Repeated patches
+sent to upstream to fix the ordering, but not yet applied.
+
+30. When jumping to a station, the position will overrun by the z-offset
+of the cab in the '#COCKPIT' section of 'train.dat'.
+
+31. When running the 'HighDetail_323_Summer_2002_0931_Rain_Overcast.csv'
+diagram of the Cross City-South route under the AI, the AI will get
+stuck at 2.5 metres before "Station 4" (a signalling-onlynon-station
+after Northfield). The AI stops outside the allowed tolerance and so
+the signal does not clear and the AI does not attempt to progress any
+closer. (1.2.1.1 onwards).
+
+32. If mouse-look is in use (click the right mouse button once) and
+Ctrl-f is pressed to switch between windowed and fullscreen, then the
+viewport direction will randomly move; leaving the camera in a number
+of strange directions (eg. upside down, sideways or looking backwards).
+More pronounced with the cab camera restriction disabled (Ctrl-r).
+
+Fixed with outstanding patches:
+-------------------------------
+14. Loading fails if 'Train/' and 'Railway/' do not share a common parent
+directory at some point. # Disabled the check for 'Train/' as it is
+not actually used later (loading works fine when a Train/ and Railway/
+are in separate locations.
+
+26. On closing, it may go into an infinite loop waiting on a futex
+(100% CPU). Currently worked around by adding a Watchdog that will
+force the program to exit.
+
+Remaining:
+----------
+1. The binary 'OpenBve.exe' binary compiled by Michelle for MS Windows
+may crash out of the box; hopefully this patched recompile doesn't!
+(Whether this is true depends on the particular release!)...
+
+9. The preview images for trains and route in the file-manager screen
+display random binary data if the preview image isn't found (normally
+from being the incorrect case). This also applies to flag icons.
+
+10. Crashes if you provide correct openbve /train=... /route=...
+
+11. Crashes if you provide incorrect openbve /train=... /route=...
+
+13. Pressing "Show issues" during a route with loads of reported issues
+causes 100% CPU.
+
+17. CPU load goes up to 100% is OpenBVE is obscured from view
+(eg. Alt-tab to another program).
+
+19. The program may sometimes display "route not found" or "train not found"
+even when the correct files do exist. clicking on another route/directory
+and then clicking back works around it, so this is probably a race condition.
+
+22. Excessive re-drawing of the listview boxes causes the interface to lag
+(particularly noticable when typing in the route selection path box.
+
+23. If the Train/ and Railway/ do not share and immediate common parent
+directory, the external view (may) not be loaded.
+
+24. Route-specific Timetable overlays try to be loaded from
+Train/somewhere/../../Railway/Object/... rather than directly from
+Railway/Object/...
+nb. "Not a bug", according to upstream, but timetables when driving
+alternative trains (as suggested/encouraged by the documentation/user
+interface) do not work otherwise...
+
+25. After a route is selected in the listview, the circular "busy"
+mouse pointer may stay busy even after loading is completely. As
+a workaround, move the mouse-pointer out of the listview rectangle
+and back in again.
+
+27. First noticed in 1.0.2.0; alpha shading under bridges and lampposts
+will flicker and may be yellow instead of gray. Other people have noticed
+similar issues since at least 0.9.3.3:
+http://openbve.freeforums.org/texture-colouration-in-xcs-t445.html
+
+28. First noticed in 1.0.2.0; the formLoading dialogue is not getting
+Diposed() (closed/destroyed) when the main Game Window opens and
+hangs around in the background without being redrawn.
+
+In-progress:
+------------
diff --git a/debian/changelog b/debian/changelog
index 900931c..9d91cfa 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -11,6 +11,7 @@ openbve (1.4.0.0-2) UNRELEASED; urgency=low
(removed >= 1.2.10.0-0ubuntu2) to stop errors
* Provide more context for licensing discussions with upstream
* Drop lynx dependency now that 'dh_installchangelogs' is used
+ * Merge 'known-issues.txt' into main 'README.Debian'
-- Paul Sladen <debian at paul.sladen.org> Fri, 13 Jan 2012 23:00:00 +0100
diff --git a/debian/known-issues.txt b/debian/known-issues.txt
index 36ef383..e69de29 100644
--- a/debian/known-issues.txt
+++ b/debian/known-issues.txt
@@ -1,165 +0,0 @@
-Known bugs/issues with running OpenBVE on Linux/Ubuntu/Debian.
-==============================================================
-
-Fixed and merged by upstream:
------------------------------
-2. If you have texture flashing issues; this is because the C-based
-OpenGL code was receiving pointers to single integers from the
-OpenBVE. Fixed by passing a array. Fixed upstream by passing
-components separately.
-
-3. Crashes out whilst trying to promisicously read non-accessible
-directories (eg. '/lost+found' or '~/.dbus'). It is hunting
-for directories containing a valid 'train.dat' description file.
-Kludged for the moment with try{}catch{} to not die.
-
-15. Window manager icon loading was not handled.
-
-5. Non-existant audio; sound should work the first time, but after
-quitting OpenAL complains about uncleaned-up buffers. You will
-likely need to 'kill -9 pulseaudio', although this may not recover
-the situation. Something in OpenAL/SDL isn't being cleanly shutdown.
-You may get the occasional sound afterwards on second/third attempts
-but not the full multi-channel surround experience.
-
-6. Captialisation. The routes have a combination of UPPER, lower, and
-MiXeD cAsE filenames which is a problem on Unix. The file-loading
-routinues need modifying to search for alterntively cased files if
-the filename as spelt can't be found.
-
-7. The London Undergrond cab 'LT_C69_77' appears in the centre of
-the screen in OpenBVE, meaning that you can't see where you're going.
-As a work-around, you may able to unlock the cam and move yourself up
-a bit to look over the top of the console. Sort-of-fixed in 0.9.5.5.
-(Though you will need to unlock the camera and zoom out to see any of
-the panel). This is to do with the camera-restriction box algorithm
-used in OpenBVE.
-
-8. The sprites Gauges in cabs displayed are not displayed correctly,
-instead showing a different/stretched part of the overlay with blue
-tinting from the edges of the alpha mask. Clipping/position broken.
-You will have to rely on the small rectangle OpenBVE overlays
-('EMG/B3/B2/B1/N/P1/P2/P3') instead.
-
-21. 'Security system warning', this refers to signalling system, not
-to the program trying to perform anything naughty! As extension DLLs
-are not being available on Linux. Patched to only display a warning
-*if* an 'ats.cfg' was present . This avoids the Issues dialogue on an
-otherwise error-free route/train.
-
-22. Hidden dot files/folders (those starting with '.') are clutter
-the filemanager. Patched to skip listing dot-files; you can
-still type the full name in. May be worth restricting to only
-!Platform.Windows.
-
-26. British English and flag should be en-GB/GB (not 'en-UK/UK').
-
-21. In full-screen on Linux, the mouse-cursor (pointer) remains visible in
-the centre of the screen.
-
-18. ExitToMainMenu causes a fork() and return (leaving the user at the
-command prompt and daemonising the second instance into the background)
-rather than using exec() or restarting itself. This crashes OpenBVE.
-It needs openbve to pass "/fhs" back to itself if it was handed it.
-
-4. Trying to write the directory the executable is in; On Unix this is
-likely to be owned by route. Modified to use the CLR SpecialFolder
-functions and save configuration data in the right places.
-
-16. After selecting an exterior camera view and then switching back to
-the interior camera; the cab/panel overlay appears in a different position.
-
-12. Language Locale not automatically set by default; patched to use
-System.Globalization.CultureInfo.CurrentCulture.Name as default, but
-needs refining to cope with 'll_CC' not just 'll' language codes.
-This is done on first-load but then forcifully saved in the options
-even if the user has not manually selected a language (prevents
-future automatic language selection working).
-
-20. No way reset the Train (and camera) to be upright after a crash. You
-should drive more carefully next time!
-
-29. When running on non-Win32 systems where Win32 plugin loading is not
-attempted, the message given may be the current one. Repeated patches
-sent to upstream to fix the ordering, but not yet applied.
-
-30. When jumping to a station, the position will overrun by the z-offset
-of the cab in the '#COCKPIT' section of 'train.dat'.
-
-31. When running the 'HighDetail_323_Summer_2002_0931_Rain_Overcast.csv'
-diagram of the Cross City-South route under the AI, the AI will get
-stuck at 2.5 metres before "Station 4" (a signalling-onlynon-station
-after Northfield). The AI stops outside the allowed tolerance and so
-the signal does not clear and the AI does not attempt to progress any
-closer. (1.2.1.1 onwards).
-
-32. If mouse-look is in use (click the right mouse button once) and
-Ctrl-f is pressed to switch between windowed and fullscreen, then the
-viewport direction will randomly move; leaving the camera in a number
-of strange directions (eg. upside down, sideways or looking backwards).
-More pronounced with the cab camera restriction disabled (Ctrl-r).
-
-Fixed with outstanding patches:
--------------------------------
-14. Loading fails if 'Train/' and 'Railway/' do not share a common parent
-directory at some point. # Disabled the check for 'Train/' as it is
-not actually used later (loading works fine when a Train/ and Railway/
-are in separate locations.
-
-26. On closing, it may go into an infinite loop waiting on a futex
-(100% CPU). Currently worked around by adding a Watchdog that will
-force the program to exit.
-
-Remaining:
-----------
-1. The binary 'OpenBve.exe' binary compiled by Michelle for MS Windows
-may crash out of the box; hopefully this patched recompile doesn't!
-(Whether this is true depends on the particular release!)...
-
-9. The preview images for trains and route in the file-manager screen
-display random binary data if the preview image isn't found (normally
-from being the incorrect case). This also applies to flag icons.
-
-10. Crashes if you provide correct openbve /train=... /route=...
-
-11. Crashes if you provide incorrect openbve /train=... /route=...
-
-13. Pressing "Show issues" during a route with loads of reported issues
-causes 100% CPU.
-
-17. CPU load goes up to 100% is OpenBVE is obscured from view
-(eg. Alt-tab to another program).
-
-19. The program may sometimes display "route not found" or "train not found"
-even when the correct files do exist. clicking on another route/directory
-and then clicking back works around it, so this is probably a race condition.
-
-22. Excessive re-drawing of the listview boxes causes the interface to lag
-(particularly noticable when typing in the route selection path box.
-
-23. If the Train/ and Railway/ do not share and immediate common parent
-directory, the external view (may) not be loaded.
-
-24. Route-specific Timetable overlays try to be loaded from
-Train/somewhere/../../Railway/Object/... rather than directly from
-Railway/Object/...
-nb. "Not a bug", according to upstream, but timetables when driving
-alternative trains (as suggested/encouraged by the documentation/user
-interface) do not work otherwise...
-
-25. After a route is selected in the listview, the circular "busy"
-mouse pointer may stay busy even after loading is completely. As
-a workaround, move the mouse-pointer out of the listview rectangle
-and back in again.
-
-27. First noticed in 1.0.2.0; alpha shading under bridges and lampposts
-will flicker and may be yellow instead of gray. Other people have noticed
-similar issues since at least 0.9.3.3:
-http://openbve.freeforums.org/texture-colouration-in-xcs-t445.html
-
-28. First noticed in 1.0.2.0; the formLoading dialogue is not getting
-Diposed() (closed/destroyed) when the main Game Window opens and
-hangs around in the background without being redrawn.
-
-In-progress:
-------------
diff --git a/debian/openbve.docs b/debian/openbve.docs
index c92c73d..5ee78af 100644
--- a/debian/openbve.docs
+++ b/debian/openbve.docs
@@ -1,4 +1,3 @@
-debian/known-issues.txt
debian/routes-and-trains.txt
debian/releasenotes.html
debian/credits.html
--
openbve
More information about the Pkg-cli-apps-commits
mailing list