Bug triage/2007-06-11

Stats

Not Resolved: 1086

• Updated Sunday: 113
  • Of those still open: 96
  • Of those being reopened: 8

Triage Queue: 838

• Blockers: 20
• Critical: 53
• Others (votes>=3): 143

Last Call items before Triage

• VWR-225 - SConstruct: create newview lib before final link - Dzonatas Sol
  • 3 votes
  • I've discussed this issue with a Linden, and it most likely will grow stale due to the nature of planned changes to the source. If we don't want to leave open (now) very low priority issues then close. Dzonatas Sol 20:44, 9 June 2007 (PDT)
  • Action: Closed
• VWR-652 - A harmless compiler warning in indra.l.cpp - Alissa Sabre
  • not sure for windows, one-liner (NB)
  • Action: Import + assign to Soft
• VWR-727 - Torn-off IM windows should be minimizable - Boroondas Gupte
  • one liner, already imported as SL-37565, seems to have been a confusion with other issue. Now patch attached. (NB)
• VWR-349 - Change keyboard shortcuts, because entering { [ ] } on German and some other international keyboards (AltGr 7, 8, 9, 0) triggers Rendering Features accelerators Ctrl-Alt-7, 8, 9, 0 (previously resulting in unstable viewer) - Thomas Shikami
  • obvious, one file, 10 lines, Had spoken about this last time, but not imported (NB)
  • I don't have a German keyboard to test this. Perhaps, we can import and ask to borrow Linden help. Dzonatas Sol 10:07, 11 June 2007 (PDT)
  • I have a German keyboard. There AltGr equals Alt+Ctrl. AltGr+9 (what Europeans type for "]") does toggle for example the Flexi render feature (verified tested here), because it technically is Alt+Ctrl+9. So the patch is simply about moving the debug render features to something that does not gets triggered through Alt+Ctrl+7/8/9/0 (because Europeans indirectly type this when typing "{", "[", "]", "}" ). I.e. if the patch let's you tigger the features with Ctrl+Alt+F9 it does what I intended it to do :-) (NB)
  • Action: Import
• VWR-1151 - Updated VC8 solution files - Nicholaz Beresford
  • Given that VC8 is still not fully supported, this will still need last minute updates even after import. Dzonatas Sol 12:14, 11 June 2007 (PDT)
  • It doesn't really "need" to be imported ... the ZIP on the JIRA will do nicely as far as the Wiki is concerned, so this is a bit like a Meta-Issue (although the worst thing that can happen is to replace a a completely outdated set of files with one that's a little less outdated :-)) (NB)
  • Action: Import + discuss at LL

Issues that have had a review pass or need further research/inquires

• VWR-1159 - SLurl do not work under Linux (No way to manage "secondfile://" protocoll under linux) - Tayra Dagostino
  • (sorry if this add is wrong, i'm new to bug triages, i'm going to attend today) (OO)
  • 4 votes (submitted 2 hours ago. this is an old, old bug)
  • [noted: no patch] Dzonatas Sol 12:05, 11 June 2007 (PDT)
  • Action: Import
• VWR-778 - Unnecessary blurry video/textures with low end graphics cards - Nicholaz Beresford
  • Reviewed, patch ok, but not tested. Dzonatas Sol 11:19, 11 June 2007 (PDT)
  • Action: Import
• VWR-446 - Automatically start renaming new user-created assets and automatically select new user-created folders - Mr Greggan
  • 7 votes
  • Patch is not ready. We'll need to ask if it is still being worked by Still Defiant, or we could import without patch. Dzonatas Sol 11:45, 11 June 2007 (PDT)
  • Action: Import
• VWR-597 - Abuse report tool should autofill abuser name when reporting an object - Tateru Nino
  • 6 votes
  • "I'd like to make sure that the AR tool does not become a way to easily exploit UUIDs; otherwise, nice patch!" Dzonatas Sol 11:59, 11 June 2007 (PDT)
  • Action: Done already
• VWR-631 - need Second Life Viewer compiled under Solaris x86 and SPARC - Whoops Babii
  • This one is either stale, or there is no movement on support for Solaris. Dzonatas Sol 00:35, 10 June 2007 (PDT)
  • Action: Import
• SVC-104 - llBase64ToInteger caveats & bugs - Strife Onizuka
  • Should be looked at, since it apparently involves a bug where uninitialized data is referenced
  • This would be helpful if it can be re-implemented as a patch. Dzonatas Sol 12:11, 11 June 2007 (PDT)
  • I'm pretty sure this is a server-side patch, not something that can be patched in the client Able Whitman 14:33, 11 June 2007 (PDT)
  • Action: Import
• VWR-190 - Replace HTML Help with Harry Linden's Google Co-op search, or add a button to the HTML Help window to go thee. - SignpostMarv Martin
  • 4 votes
  • This should be a temporary fix if desired, as this may change. Perhaps even if planned changes occur, the button can co-exist, but it is unlikely to be as helpful. Dzonatas Sol 12:24, 11 June 2007 (PDT)
  • Action: Import + assign to Harry

Not Ready for Triage

• VWR-654 - Tools / Reset all Scripts in Selection Window does not Scroll automatically - Zi Ree
  • The first patch, while helpful, is just a simple one, and it needs a follow-up patch to complete it. Dzonatas Sol 09:38, 11 June 2007 (PDT)
• VWR-733 - Change Request: Excess texture buffer memory size resulting in disk thrashing/crashes/virt memory swapping - Nicholaz Beresford
  • Reviewed, the patch needs clean-up. Dzonatas Sol 11:19, 11 June 2007 (PDT)
• VWR-893 - Patch to make View Admin Options sticky - Gigs Taggart
  • Reviewed, an added llcontroldef does not appear to be the best solution for this feature. Dzonatas Sol 11:54, 11 June 2007 (PDT)

Transcript

[15:02]	Rob Linden:	should we get started?
[15:03]	Stevex Janus:	ok I
[15:03]	Rob Linden:	VWR-225
[15:03]	Wyn Galbraith:	Greeting to Bridie too.
[15:03]	Bridie Linden:	Guess I should put away my snowball thrower now
[15:03]	Bridie Linden:	sorry for being late
[15:04]	Wyn Galbraith:	Why?
[15:04]	Wyn Galbraith:	I mean why put it away ;)
[15:04]	Soft Linden:	What meeting's not improved by snowballs?
[15:04]	Bridie Linden:	good pts
[15:04]	Wyn Galbraith:	Or Rubber bands.
[15:04]	Able Whitman eyes his beehive launcher
[15:05]	Soft Linden:	Was it the rubberbands or the snowballs that were VWR-225?
[15:05]	Saijanai Kuhn:	special lesson learned in USAF: don't shoot the captain when he walks by...
[15:05]	Rob Linden:	is Dzonatas here by any chance?
[15:05]	Rob Linden:	shoot...nope
[15:05]	Nicholaz Beresford:	doesn't seem so
[15:05]	Saijanai Kuhn:	they teach officers to stop rubber band fights with a signle glance. Special class
[15:06]	Rob Linden:	I'm not entirely sure what to do with VWR-225 ....
[15:06]	Nicholaz Beresford:	I don't even know what Sconstuct is :-)
[15:06]	Soft Linden:	Might be worth holding until Dz is here. From the look of it, it's not intended to be integrated, only to offer a benchmark?
[15:06]	Rob Linden:	k... postponed
[15:06]	Soft Linden:	(From the 2nd to last comment)
[15:07]	Rob Linden:	VWR-652
[15:08]	Nicholaz Beresford:	dunno what it does. have tried it on windows but didn't appear to make a difference (none that I could notice)
[15:09]	Able Whitman:	Yeah, I never got the warning under VC8, but the patch looks safe enough.
[15:09]	Rob Linden:	mucking with yacc files is always a little precarious, from what I understand
[15:09]	Rob Linden:	they tend to be pretty fragile
[15:09]	Wyn Galbraith:	Lag test
[15:10]	Wyn Galbraith phews and is good.
[15:10]	Nicholaz Beresford:	given that it's Alissa ... maybe with an #ifdef MAC
[15:10]	Multi Gadget v1.52.0 by Timeless Prototype
[15:10]	Able Whitman:	Well, the bug mentions that the warning is harmless, so if yacc files are fragile, not taking the patch at least doesn't make things any worse..
[15:11]	Soft Linden:	Cursory googling suggests it has no effect on the output in modern versions of yacc. It wouldn't be difficult to do a quick A/B test to verify output didn't change.
[15:11]	Rob Linden:	btw....I'll be posting the transcript for this
[15:11]	Nicholaz Beresford:	k
[15:11]	Soft Linden:	Hey, Dzonatas :)
[15:11]	Dzonatas Sol:	hello =)
[15:11]	Wyn Galbraith:	Speak of the Devil... :)
[15:12]	Rob Linden:	ok....I'll import it
[15:12]	Rob Linden:	let's loop back to VWR-225 now that Dz is here
[15:12]	Soft Linden:	Go ahead and assign it directly to me if you like. Should be a quickie.
[15:12]	Dzonatas Sol:	225 a quickie?
[15:13]	Soft Linden:	Sorry - no, VWR-652 - the prior one.
[15:13]	Dzonatas Sol:	k k
[15:13]	Rob Linden:	VWR-225 : Dzonatas...that's the SConstruct changes you made. what would you recommend at this poing?
[15:13]	Rob Linden:	point?
[15:14]	Rob Linden:	(heh...I kinda like "poing")
[15:14]	Dzonatas Sol:	i rather see a vote... but...
[15:14]	Dzonatas Sol:	i hope that stale issues , low priority... can be like sit around for a few months if no one touches... and...
[15:15]	Kamilion Schnook:	Quick question, rob; Can we get a "Email" component in Jira's SVC listing? I was just heading there to post my email issue.
[15:15]	Dzonatas Sol:	maybe give it a second chane on the import.. note that it is stale... if its sits longer internally... completly close
[15:15]	Soft Linden:	From the 2nd to last comment on that issue - is it meant to be imported, or is it only for illustration of the speed difference?
[15:15]	Dzonatas Sol:	so like it would have a total of a half year on stale issues before we just close
[15:16]	Soft Linden:	The "this patch alone cannot fix the issue. It only illustrates the difference in compile time"
[15:16]	Dzonatas Sol:	it can't be used as is
[15:16]	Dzonatas Sol:	it only shows the difference in compile time
[15:17]	Rob Linden:	ok....hmmm. how much work would it be to make it ready for application?
[15:17]	Dzonatas Sol:	many globals would have to be re-organized for them to initialize correctly.
[15:17]	Dzonatas Sol:	there is already talk about a reorg work on the source, so the issue may be moot at that time
[15:18]	Rob Linden:	ah, I see
[15:19]	Dzonatas Sol:	i don't waste time with it... I'll just close it
[15:19]	Dzonatas Sol:	next...
[15:19]	Rob Linden:	k....VWR-349
[15:20]	Soft Linden:	Dzonatas' issue is linked against an internal issue suggesting that split. So it's still there as reference if someone takes that on.
[15:20]	Rob Linden:	we discussed this one before I think. I must have forgot to import
[15:20]	Dzonatas Sol noted
[15:21]	Nicholaz Beresford:	yes, we spoke about 349 last time.
[15:22]	Rob Linden:	k...next, VWR-727 (accidently skipped)
[15:23]	Nicholaz Beresford:	pretty obvious.
[15:24]	Soft Linden:	And here I'd forgotten those even tore off. Do people use that a lot?
[15:24]	Nicholaz Beresford:	I have no idea :-)
[15:24]	Dzonatas Sol:	tear-off and minimize... acts as a good filter
[15:24]	Able Whitman:	I never do, but I can see how the torn-off windows not being minimizable might be pretty annoying.
[15:24]	Bridie Linden:	nice to have!
[15:25]	Soft Linden:	Sure. If tear-off is there, should have minimize like any other window
[15:25]	Nicholaz Beresford:	can't see how it would hurt either.
[15:25]	Rob Linden:	hmmmm....it creates a problem when one person submits for someone else. did Boroondas contribute it to Linden Lab?
[15:25]	Kamilion Schnook:	I use the tear off occasionally with the Client - Rendering - Types menu.
[15:25]	Nicholaz Beresford:	for a one-liner in a xml file?
[15:25]	Rob Linden:	oh...is that all it is?
[15:25]	Rob Linden:	heh...whaddya know
[15:26]	Rob Linden:	ok....sure, why not
[15:26]	Rob Linden:	import
[15:26]	Dzonatas Sol:	import already done by other issue
[15:26]	Nicholaz Beresford:	btw, yes he asked me in email to attach the file because he couldn't.
[15:28]	Rob Linden:	ok...wasthere a reason for putting it on the agenda?
[15:28]	Rob Linden:	(next up: VWR-1151)
[15:29]	Nicholaz Beresford:	pretty simple. the _vc8 files in the archive are currently useless.
[15:29]	Nicholaz Beresford:	either take them out or take new ones in.
[15:29]	Able Whitman:	As a user of VC8, I'd like to avoid having to re-convert the project files every release, so I like fixing this bug a lot :)
[15:30]	Stevex Janus:	!Is there a way to generate them automatically?
[15:30]	Dzonatas Sol:	VC8 + scons is much nicer
[15:30]	Dzonatas Sol:	=)
[15:30]	Nicholaz Beresford:	make a wiki about it and tell people how it works :-)
[15:30]	linux soda machine Mark II v1.05: Menu timed out.
[15:31]	Dzonatas Sol noted
[15:31]	Nicholaz Beresford:	but basically, having the vc8 set on the jira serves 90% of the purpose anyway.
[15:31]	Soft Linden:	VC8 does parallel builds on all cores, which is already a huge step up. Does the auto-upgraded project work okay? Or does it take some tweaking?
[15:31]	linux soda machine Mark II v1.05: Here's your can of Red Hat Fedora flavored Linux soda, enjoy!
[15:32]	Nicholaz Beresford:	as far as I can tell it works out of the box on 1.17.0.9 but I don't know what the release will look like.
[15:32]	Able Whitman:	It takes some tweaking, mostly due to the need for additional symbol defines to enable depricated CRT functions and rever to the old w_char behavior.
[15:32]	Rob Linden:	I think there's a conversation that we need to have at LL about this, but I think what I'm going to do is import this, and use the imported issue as a centerpice for that conversation
[15:33]	Kamilion Schnook grins
[15:33]	Nicholaz Beresford:	(sorry, forget my last comment ... was a misunderstanding. the auto-upgrade takes some work)
[15:33]	Able Whitman:	Hey, if you folks all just moved to VS2005, we wouldn't have to worry about this any more :)
[15:33]	Dzonatas Sol points to a version of distcc that works with msvc
[15:33]	Rob Linden:	next: VWR-1159
[15:34]	Soft Linden:	I think the VS2005 upgrade is limited by some 3rd party server libs that don't link in 2005 yet.
[15:35]	Nicholaz Beresford:	can't say anythig about 1159. I'm on Windows :-)
[15:35]	Dzonatas Sol:	good, has repro
[15:35]	Able Whitman:	Judging from the comments on 1159, it looks like the actual issue is broader than broken SLURL handling...
[15:36]	Stevex Janus:	Ah so it is not just me not being able to use SLurls
[15:36]	Kamilion Schnook:	Yeah, I've also noticed the linux viewer allows running multiple instances without -multiple on the commandline.
[15:36]	Opensource Obscure:	as a linux user i'd say that fixing slurl only, maybe with a workaround, would be a great improvement itself.
[15:37]	Saijanai Kuhn:	they don't open the map in MacOS BTW, if they are supposed to
[15:37]	Rob Linden:	can someone fix the typo in the subject line
[15:37]	Rob Linden:	("secondfile")
[15:37]	Kamilion Schnook:	on it.
[15:38]	Laetizia Coronet:	the SLurl site? No map in Windows/Firefox either - correct me if I am mistaking the issue
[15:38]	Able Whitman:	Hmmm... It would be helpful to know if multiple instance detection and SLURLs are both broken under Linux and Mac, and then have separate bugs for those specific issues.
[15:39]	Opensource Obscure:	btw multiple instances work correctly on linux using -multiple .
[15:39]	Able Whitman:	I think the fact that the map on slurl.com is blank is a known issue, unrelated to 1159.
[15:39]	Laetizia Coronet:	ok
[15:39]	Kamilion Schnook:	VWR-1159 rename completed.
[15:39]	Able Whitman:	Yes, but the comments for 1159 say that the client opens multiple instances even without the -multiple switch, which is the wrong behavior.
[15:40]	Soft Linden:	Is there a standard way to register protocol schemes with linux... secondlife:// et al, or is it handled by browser?
[15:40]	Rob Linden:	there's several standards :)
[15:40]	Stevex Janus:	brwose
[15:40]	Kamilion Schnook:	GNOME and KDE can both handle new mime types otherwise it's handled in the browser directly.
[15:40]	Soft Linden:	Sounds like one bug is detecting another running instance, and the other is verifying the scheme is handled by the popular browsers
[15:41]	Rob Linden:	should we break it up before import?
[15:41]	Kamilion Schnook:	AFAIK firefox will also let you register a protocol handler/mime type directly, which should work on all three platforms.
[15:41]	Soft Linden:	sure... real quick hunting for dupes on those
[15:42]	Rob Linden:	so, VWR-1159 should be the multi-instance bug, I think.
[15:42]	Opensource Obscure:	soft: the second bug finds a workaround in tweaking the browser settings. users could be directed to their browsers faqs
[15:43]	Kamilion Schnook:	Too bad JIRA doesn't have bug dependancys; that might come in handy for these cases ;)
[15:43]	Rob Linden:	Kamilion...actually, it does
[15:43]	Kamilion Schnook:	In that case, why not create a new JIRA for the -multiple issue and link 1159 as a child issue of that, since it's explicit in being linux.
[15:43]	Soft Linden:	I don't find either internally
[15:44]	Rob Linden:	k...I'll import VWR-1159. could someone copy the transcript here into the comments?
[15:44]	Kamilion Schnook:	How far back?
[15:45]	Rob Linden:	where we first started talking about it
[15:45]	Able Whitman:	My apologies, my net connection went away for a moment.
[15:45]	Kamilion Schnook:	Done.
[15:45]	Rob Linden:	we should be able to import whatever dep bugs people file based on this conversation later
[15:46]	Rob Linden:	next up, VWR-778
[15:46]	Fremont Cunningham:	Rob, are you dealing with issues in cronological order, or off some pre-decided agenda, or what?
[15:46]	Nicholaz Beresford:	basically it means you don't get a clear picture on a PC with 16MB graphics card
[15:47]	Kamilion Schnook:	https://wiki.secondlife.com/wiki/Bug_triage/Agenda
[15:47]	Stevex Janus:	Frem, its on the wiki
[15:48]	Fremont Cunningham:	Which wiki?
[15:48]	Rob Linden:	looks reasonable. will import
[15:48]	Kamilion Schnook:	Link I just posted, Fremont.
[15:48]	Nicholaz Beresford:	I have seen it (results from 778) on an office laptop.
[15:49]	Wyn Galbraith:	What OS is that for?
[15:49]	Soft Linden:	Sure, patch looks straightforward. Puts a minimum size on the system memory buffer. Hopefully someone endeavoring to run the viewer with a <64M card is compensating with some extra system memory :/
[15:49]	Rob Linden:	next up: VWR-446
[15:49]	Nicholaz Beresford:	soft: my gf has an IBM with 16MB graphics card and 1GB main mem :-)
[15:50]	Squirrel Wood:	humm.. even on my nvidia 7950 GT I get to see blurry textures more often than I like
[15:50]	Kamilion Schnook:	My laptop uses a selectable amount of system memory as graphics memory. 8/16/32MB.
[15:51]	Nicholaz Beresford:	you'll even see the difference between 32 and 64MB there.
[15:51]	Nicholaz Beresford:	446 makes a lot of sense to me ... can't speak for the patch though.
[15:51]	Dzonatas Sol:	i didn't get the chance to IM Still
[15:51]	Dzonatas Sol:	doing now...
[15:52]	Kamilion Schnook:	Normally I keep it on 8MB because this thing's crappy ATI Radeon Mobility U1/IGP320M can't do hardware texture and lighting
[15:52]	Rob Linden:	I'll import it. looks like a good idea, even if the patch is so-so.
[15:53]	Rob Linden:	I'll let whoever incorporates the patch make the call
[15:53]	Rob Linden:	next up: VWR-597
[15:53]	Kamilion Schnook:	seen here: http://www.sllabs.com/kamilion/slwireframe.png
[15:53]	Able Whitman:	I like the idea of 446 as well, but since it touches the finicky inventory system, I hope it gets decent regression testing :)
[15:54]	Soft Linden:	597 says it's already applied
[15:54]	Rob Linden:	yup
[15:54]	Bridie Linden:	/nods
[15:54]	Rob Linden:	next: VWR-631
[15:54]	Bridie Linden nods
[15:54]	Nicholaz Beresford:	that must be pretty new ( fixed)
[15:55]	Squirrel Wood:	hmm.. last time I tried to abuse report someones objects search was not finding the abusers name and thus, filing the report was not possible :p
[15:55]	Rob Linden:	I'll import VWR-631....we've been talkking to Whoops offline
[15:56]	Kamilion Schnook:	Good, another supported platform.
[15:56]	Rob Linden:	next up: SVC-104
[15:57]	Nicholaz Beresford:	I didn't really understand what it's referring to, the patch doesn't seem to apply to something in the viewer .... at least I couldn't find it
[15:57]	Rob Linden:	oh....yeah, that's weird
[15:57]	Dzonatas Sol:	the patch is a test case
[15:57]	Kamilion Schnook:	It doesn't, it's a example he says.
[15:58]	Rob Linden:	ah...right.
[15:58]	Able Whitman:	No, the bug refers to the behavior of the llBase64ToInteger() function, which seems to not initialize the buffer it uses to parse the base64 string
[15:58]	Kamilion Schnook:	I belive it's for the serverside LSLVM
[15:58]	Rob Linden:	k...will import
[15:59]	Able Whitman:	I think it's good to fix accesses to uninitialized memory, but I also think the attached suggested fix is less than ideal.
[15:59]	Rob Linden:	next: VWR-190
[15:59]	Dzonatas Sol:	boost lib has a conversion, likewise
[15:59]	Able Whitman:	Ah, then use the boost version :)
[16:00]	Wyn Galbraith has to run.
[16:00]	Kamilion Schnook:	190 sounds good to me.
[16:00]	Able Whitman:	Yeah, the Google SL portal is very handy
[16:01]	Wyn Galbraith:	Thanks Rob, and Bridie and Soft.
[16:01]	Dzonatas Sol:	ttyl Wyn!
[16:01]	Bridie Linden:	cya!
[16:01]	Soft Linden:	Too late
[16:01]	Soft Linden:	Might import that and assign Harry to ask if it's a good replacement
[16:01]	Kamilion Schnook:	Ah well, it'll be in the transcript ;)
[16:01]	Rob Linden:	soft: good idea
[16:02]	Rob Linden:	ok...I'm timing out now
[16:02]	Rob Linden:	thanks everyone for showing up!
[16:02]	Dzonatas Sol:	woot! got through the list =)
[16:02]	Rob Linden:	w00t indeed
[16:02]	Nicholaz Beresford:	kewl
[16:02]	Dzonatas Sol:	take care all
[16:02]	Dzonatas Sol:	thank for being here
[16:03]	Able Whitman:	Thanks, Rob, Soft, and Bridie!
[16:03]	Soft Linden:	Okie. Back to work...
[16:03]	Soft Linden:	Thanks for coming, all. Thanks for hosting, Rob!