| Age | Commit message (Collapse) | Author |
|---|
|
for PATH-542. The viewer calculated rotation resulting from an object's angular velocity was being incorrectly reset when the corresponding object update was received. With this bugfix, the rotation resulting from the angular velocity is accumulated separately and is re-appplied when the object update resets the object's rotation.
|
|
For as long as I've been paying attention to viewer logs, every run has always
stated -- in all-caps, no less! :
NOTE: ALL NOTIFICATIONS THAT OCCUR WILL GET ADDED TO IGNORE LIST FOR LATER RUNS.
I have always found that unsettling, having little desire to automatically
ignore notifications -- especially not whatever random set happens to manifest
during a particular run. However, over time I've learned that it seems to have
no actual impact. I've wondered (not too hard) what it actually *does* mean --
but generally I'm looking at a viewer log because I'm in hot pursuit of some
quite different problem.
Stumbling across the source for this message today in
LLViewerWindow::LLViewerWindow, I was shocked -- shocked! -- to find it was
being produced unconditionally! Therefore it has heretofore been utterly
meaningless log spam. It provided no additional information whatsoever to
anyone reading a log -- only that vague sense of unease.
Make this conditional on gSavedSettings.getBOOL("IgnoreAllNotifications"),
which would appear to be the original intent. Now its presence (or absence)
actually says something about internal conditions in the viewer.
|
|
Popup text used to end with the question:
Visit [[create_account_url] [SECOND_LIFE] web site] to create a new account?
presumably to set context for the buttons labeled "New Account..." and
"Continue". At Leo's request, remove the question and relabel the yes button
"Create Account...".
|
|
Login-panel logic distinguishes "system grid" from "non-system grid." With
Oz's recent changes for pathfinding, now only agni and aditi are "system
grids;" anything else configured into grids.xml is a "non-system grid." The
difference is that when you select a "non-system grid" on the grid selector,
we turn off the "lost password?" link and the "create account" button -- since
how can we help with either if we don't recognize the grid? This logic already
existed, but only turned off the create-account button, leaving the new title
"CREATE YOUR ACCOUNT" over an empty corner of the login panel. Turn that off too.
|
|
|
|
Some viewer-development code had been moved, and so wasn't patched with my
panel_login layout changes; verified each of my llpanellogin.cpp commits
against new tip rev. Reformatted panel_login.xml in the spirit of the
preferred indentation scheme but with my layout changes.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
cancel), the pathing console rebuilds the gfx data.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
(259197) Beta in "Add scroll list item"
removed linear scan for namelist items and used weak reference (LLHandle) instead
|
|
|
|
|
|
|
|
and characters floaters.
|
|
|
|
|
|
Although the XUI for this popup, and the code to display and respond to it,
have been present from the first rev in the current viewer repository, I can
find no trace of it ever having been called.
Reword embedded link in popup message in each supported language: directly
mentioning secondlife.com in display text engages anti-URL-spoofing logic,
causing the popup to display the whole messy URL with sourceid and lang
params. Change to say "Second Life web site" instead -- needs translations.
|
|
Otherwise later log fields start showing up with zero fill, etc.
|
|
|
|
|
|
texture fetching queue ONLY once during a teleporting.
|
|
getUniqueID() was logging six somewhat random bytes as garbage characters.
Change to produce a hex string instead.
|
|
|
|
|
|
Having multiple dynamic width columns apparently prevents other columns from being manually resized by the user.
|
|
owner names.
|
|
The changeset above touched every consumer of the two LLRegistrySingletons
originally defined with std::type_info* as keys. Those two
LLRegistrySingletons were changed to use const char* as keys, then all
consumers were changed to pass std::type_info::name() instead of the plain
std::type_info* pointer -- to deal with the observed fact that on Linux, a
given type might produce different std::type_info* pointers in different load
modules. Since then, Richard turned up the fascinating fact that at least some
implementations of gcc's std::type_info::before() method already accommodate
this peculiarity. It seems worth backing out the (dismayingly pervasive)
change to see if properly using std::type_info::before() as the map comparator
will work just as well, with conceptually simpler source code.
This backout is transitional: we don't expect things to build/run properly
until we've cherry-picked certain other pertinent changes.
|
|
The "Start at:" combo box and the grid-selector box can be turned on and off
with checkboxes on the Preferences floater. Break the bottom-left layout_panel
on panel_login into several different layout_panels so when these UI elements
disappear, other controls move left to fill the holes. Similarly, when you
turn on either or both of the UI elements in question, the remaining controls
in that general bottom-left area move right to accommodate.
Tweak C++ and non-English panel_login.xml overrides to reflect widgets'
containment in new layout_panel elements.
Discard obsolete "mode_selection_text" and "mode_combo" translations from non-
English panel_login.xml overrides.
|
|
|
|
Revert to corresponding original login_help text for each affected language.
|
|
values change.
|
|
|
|
|
|
|
|
|
|
|
|
|
