-
TwinSteel
Did the latest stable fix the Publish window not closing after publishing ? Yes, I've unticked the box.
-
njsg
I suppose the screenshot from
bug 1112694 illustrates that problem as well?
-
njsg
"all files published" and the checkbox is not checked
-
WG9s
something like Do you wish to end the hold and resume the schedule?
-
WG9s
this is a taking questions that have a yes/no answer but then displahying buttons that say cancel and conform or any other working should just be yes or no.
-
WG9s
oops thos last 2 posassts ers intended to be in the conversation with njsg
-
IanN
-
frg
Hi tonymec
-
tonymec
hi frg
-
IanN
hi tonymec, how were the storms for you?
-
frg
next one due tonight I think
-
tonymec
hi IanN, with a heavy coat I could stand them, and you?
-
WG9s
Hi everyone here for the meeting
-
jardsonto
Hello!
-
tonymec
hi lonely one here for the meeting :-P
-
frg
Hi jardsonto
-
frg
Hi WaltS48
-
WaltS48
Hi frg, WG9s, IanN and all.
-
frg
tonymec found it not so bad. Coast was hit harder but nothing I haven't seen before. Was a bit surpised that this one was categorized as the worst since 2007.
-
WG9s
hi WaltS48
-
tonymec
frg: hm, maybe we weren't outside in the worst of it.
-
IanN
hi WaltS48
-
IanN
tonymec: got soaked
-
frg
IanN no hiking in the woods I assume :)
-
IanN
frg: not this weekend
-
tonymec
IanN: here in Brussels I felt more wind than rain.
-
frg
IanN looking at the forcast does not look good for GB the whole week depending where you live.
-
frg
west coast might get it again harder.
-
WG9s
here where I live in massachusetts sometimes storms that don;t seem that bad to me are thought to be the worst becuase of the number of beachfront proerties that got totalled. but then some of these seem to be build between where the lighthose is and the waters edge so seems to me the fault of any one who build bought approved a mortgage or insured issue as being just terminally stupid.
-
frg
WG9s they started to build houses in known natural areas subject to flooding here and then guess what happend when the flimsy levees they added to the areas broke...
-
WG9s
but really building closer to the wate rthan the lighthouse? that just seems like recipe for disaster.
-
IanN
-
tonymec
WG9s: depends... I know places with a lighthouse at the back of the harbour entrance channel and another one higher and further back, blinking isophase to create an alignment.
-
IanN
hi rsx11m
-
rsx11m
hi IanN frg tonymec WG9s
-
tonymec
hi rsx11m
-
frg
Hi rsx11m
-
njsg
hi .*
-
WG9s
It is not everywhere in the sate. People on the NOrth Shore have lost homes that were there forever, but we have these pople on our south shore who had money could pay off people to get permission to build where they should not have been and got what they deserved.
-
rsx11m
hi njsg
-
WG9s
hi rsx11m
-
IanN
-
IanN
Who's taking minutes?
-
frg
me and probably late again
-
frg
Almost forgot last time because didn't add it to my virtual post-it.
-
WG9s
so you mean late posting
-
IanN
Nominees for Friends of the Fish Tank?
-
tonymec
WG9s: in the Netherlands there are several "verdronken landen" — areas of several villages each, which were drowned by tidal storms in the Middle Ages and are still under water.
-
frg
Nominate WG9s for building our prereleases and central and keeping up with my/our doings with the patch queues.
-
tonymec
second WG9s
-
WG9s
Well thank you
-
IanN
3rded
-
IanN
Action Items
-
frg
Nothing new from me.
-
WG9s
I am still remiss in not getting my promised logbot working
-
IanN
Status of the SeaMonkey Infrastructure
-
WG9s
i do have the disk frg provided installed and it is bing backed up now so just neet to catully do the part where it logs in and logs
-
frg
All building I think. I worked a bit on further stylo updates for later rust versions than 1.37 but got bogged down quickly with rust dependencies.
-
frg
Rest assured will never be my favorite language.
-
IanN
nor rust :P
-
WG9s
hmm need to update the reomving 32-bit support to say 2022 rahter than 2021
-
rsx11m
... didn't happen in 2021, apparently ;-)
-
frg
Windows x86 still builds fine and so does Linux for now. Can keep it but don't expect me to give much support for it :)
-
frg
Maybe ewong can get some numbers from the update server?
-
IanN
we are hoping to enable heptapod hg repos soon
-
IanN
Status of the SeaMonkey Source Tree
-
frg
For the infra I might update the VS2019 for the final release. They fixed a compiler bug in it and seems to do fine.
-
frg
VS2022 17.1 is usuable too now but I intend to keep VS2019 a while longer.
-
IanN
frg: sorry jumped the gun
-
frg
For source tree have some work in progress stylo updates for 2.57 but nothing more for it. Still backburner. 2.53 is progressing nicely.
-
WG9s
my builds still use vs2019 becuase i am still keeping vs2017 around just in case need to do something with a new SeaMonkey 2.49 security update and keeping 3 version of VS around is a bit much.
-
WG9s
also 32-boit supoort I might be able to find time to keep them at least buildsng if frg gives up on them.
-
frg
WG9s well basically you use clang and the VS2019 tools only :)
-
WG9s
yes
-
frg
But VS2019 can be used with 2.49. You need to install the VS2017 toolset and XP dependency stuff.
-
WG9s
MOzilla no longer builds uisng MSVC and so I d not either. even for 2.53
-
frg
Yes Thought about switching but I like to have two compilers around for testing and debugging.
-
WG9s
plus otheres might want to build and debug using VisualStudio
-
WG9s
in the gui
-
frg
I think it is supported now with clang too because MS bundles it.
-
IanN
Release Train
-
frg
2.53.11 is late but building should begin today. Checking in the patches right now.
-
IanN
2.53.11 should be released in the next week
-
WG9s
If any latge breaking patches show up let me know and I can include in my builds
-
WG9s
late breaking
-
frg
WG9s don't think so. A few sec fixes and also found a few crash fixes for mail which are included. Already in your builds to see if they are bad. No complaints so in they go.
-
frg
Also the TrimStart and TrimEnd string properties discussed two days ago.
-
WG9s
yes was asking if something shows up not yet in mhy builds feel free to aks for me to do an eraly build if required
-
frg
WG9s nope all in and expect anything else.
-
frg
fyi gitlab 2.53.11 final branch is now updated.
-
frg
So the build fun can start
-
IanN
Extensions Tracking
-
frg
I removed a few more NPAPI bits in 2.53.12b1 pre and also started to take out gtk2 which is dependent on this.
-
IanN
github-wc-polyfill is now at 1.2.13
-
frg
Closed
Bug 1367257. If anyone objects reopen :)
-
frg
github still has problems with it but seem to be css related. gitlab seems to do better and heptapod is now supported too but didn't check.
-
WG9s
i like that I worked wth the redhat people on trying to get firefox working with gtk3. I have no idea why we would need to still support gtk2.
-
frg
WG9s we don't but the NPAPI stuff is still in and needs it for building.
-
IanN
2.Next, Feature List, Planning and Roundtable
-
frg
Bug 1675349 and
Bug 1682030 are needed. these are 40 parts and there are only so much hours in the day :)
-
rsx11m
no worries ...
-
frg
Tried just putting
Bug 1377445 and followups in but this breaks.
-
WG9s
for roundtable I am still working on geting an irc logbot and server set up. just seems not enough hours in the day!
-
frg
I got a new job starting in May so this will impact my availablity unfortunately.
-
IanN
BAU for me
-
WG9s
i do have a new drive set up for the data and mhy backup set up to back it up so now just get the config files working and things being logged and having a mechanism to make one of the channels I want to backup be private with logins required.
-
WG9s
bjt I could start with out backing that chennl up
-
frg
Doing robin round with backports. Now on microtasks but not sure if I can put them in fast.
-
WG9s
problem with the one channel trying to keep private is trying to design a way to permit people via the webserver to change their password
-
frg
But 2.53 is now beyond 57 level with base fixes and while this still sounds outdated the big reshuffles in the source done between 56 and 57 make applying follow-up patches easier.
-
njsg
I had a look at some Composer issues, I hope I can do something to add error handling and error messages to the save code; there's at least one bug about slowdown in source view but that seems to be from the editor element, when it is filled with a long-enough multiline text
-
IanN
AOB?
-
WG9s
not here
-
tonymec
neither nere
-
IanN
okay, next meeting in 2 weeks time, same bat time, same bat channel
-
rsx11m
see you in March!
-
rsx11m
bye
-
tonymec
bye
-
WG9s
bye
-
WaltS48
see ya
-
IanN
bye rsx11m and tonymec|away
-
IanN
bye walts48
-
frg
see you
-
wyatt8740
hey, I wanted to say i found a minor problem with the README that gets generated for seamonkey releases
-
wyatt8740
it says the command 'git clone -b 2_53_10_2_final
gitlab.com/seamonkey-project/seamonkey-2.53.10-comm.git comm-25310' was used to get the sources for 2.53.10, but that git repo doesn't exist
-
wyatt8740
(or is not public)
-
wyatt8740
-
wyatt8740
in other words 's/\.10//'
-
wyatt8740
-
buc
Well, the issue with "user-menu" with github-ws-polyfill-1.2.12 and SeaMonkey >= 2.53.10 seems to be introduced after the backport of
bug 1391154
-
buc
-
buc
Any thoughts are appreiated.
-
buc
tomman: If you wisely have not begun to bisect yet, then do not do it. :)
-
buc
After reverting
bug 1391154, the "user menu" works under 2.53.11b1 (with both versions of the github-wc-polyfill addon 1.2.12 and 1.2.13)
-
tomman
buc: haven't had time to figure out where I do even start, but nice to hear that there is actually a regression
-
tomman
or something
-
tomman
"This matches Chrome's implementation and the tests."
-
tomman
once again, I died inside
-
buc
tomman: They seem changed algorithm while going to use IndexOfFirstElementGt(), need to check a bit more to explain.
-
frg_Away
Hmm even with the patch the test seems flaky but less flaky:
jsfiddle.net/8f9whxbk/11
-
frg_Away
Oh wrong test. belongs to
Bug 1505471 which is not in.
-
wyatt8740
i'm sure this is a common question, but i checked out the source repos and am trying to build seamonkey 2.53.10 from them. I am running './mozilla/mach build' but it doesn't seem to be using my mozconfig (which has ac_add_options --disable-pulseaudio specified).
-
wyatt8740
additionally config.log is going in the mozilla/ subdir's object path instead of the MOZ_OBJDIR specified in the mozconfig
-
wyatt8740
do i need to be doing something else first?
-
wyatt8740
or do i just have my dirs inverted
-
frg_Away
wyatt8740 Yes inverted. The mozilla source is now the top source dir. The suite/mail code is now in comm under it. so you just use mach build. The final 2.53.11 code is now in gitlab and you can check out this branch. Lots of internal fixes/backports over 2.53.10.2
-
frg_Away
buc I just added 1505471-68a1.patch to gitlab. The reported coordinates are now correct but I doubt it will help with the problem. Worth a try if the menu uses an intersection observer and relies on the coordinates. But probably not because not an iframe.
-
buc
Well, commenting out "if (threshold == 0) threshold = -1;" in the bug's 1391154 patch resolves this issue.
-
buc
The correspond comment in the patch says: "Per the spec, we should leave threshold at 0 and distinguish between "less than all thresholds and intersecting" and "not intersecting" (queuing observer entries as both cases come to pass). However, neither Chrome nor the WPT tests expect this behavior, so treat these two cases as one."
-
buc
So, if we start to treat it as "per the spec" (instead of per Chrome), the issue gone.
-
buc
There are no any changes here in the current Firefox code tree (but this mean nothing - the current Fx development has long ceased to be the ultimate truth).
-
frg_Away
buc yeah but if I read it right it is done to match Chromes behaviour: [User impact if declined]: IntersectionObservers' behavior would not match Chrome's.
-
frg_Away
See comment 47
-
buc
frg: Unfortunately,
bug 1505471 itself does not help the issue.
-
frg
buc thought so. Was worth a try. I think I remove the per the spec line as you suggested in a separate TOP patch so that it works for now.
-
frg
a bit late for 2.53.11 unfortunately.
-
frg_Away
buc as a last shot can you try with 1711050-nsrefreshdriveroptimize-2538.patch removed.
-
buc
2.53.8 < 2.53.10 ... :)
-
frg_Away
buc I know. But just want to make sure that this is not causing it and that it worked by accident or because of the unfixed bug before. Should not be a chrome uri thou.
-
frg_Away
is this testable without needing to log in?
-
buc
No. :(
-
frg_Away
1.2.14b1 fixed the load diff css problem btw.
-
frg_Away
Added TOP-9999999-fixgithubpolyfill-25312.patch to gitlab 2.53. Will file a bug against 2.53.12 / our mozilla branch if we do not find a better solution.
-
buc
OK, LGTM
-
wyatt8740
frg_away: thanks
-
wyatt8740
I'm testing building for ppc32 linux though so I figured a release might be a little more trustworthy when trying to iron out problems
-
wyatt8740
haven't done one of these cross builds since 2018
-
frg_Away
buc thanks for tracking it down.
-
frg_Away
wyatt8740 is rust available for this arch?
-
wyatt8740
frg_away: yeah, RUSTFLAGS=--target=powerpc-unknown-linux-gnu
-
wyatt8740
the showstopper for firefox itself is node.js/V8 in the build system
-
IanN_Away
frg_Away: is rust available on OS/2?
-
frg_Away
IanN_Away no. I still wish they would release the OS/2 source but probably lost. Decommisoned my main OS/2 system and moved it into a vm. But still have it on a T42p.
-
wyatt8740
rustc --print target-list
-
wyatt8740
ooh, talking os/2? I have boxed shrinkwrapped CD copies of the 3.0 connect bundle, both red and blue spines
-
wyatt8740
i think the only machine i might be able to get it running on is my thinkpad 350 or _maybe_ a latitude D610
-
wyatt8740
due to overall lack of experience
-
frg_Away
IBM s*cks. The managers are the Dinos not the older employees and should be made redundant. They ruined everything. PS/2 MCA, token ring, Thinkpads, DB2 and now on the last leg with z/OS. Seeing migration projects left and right here. Could possibly make a fortune if I would turn myself into a consultant for these.
-
wyatt8740
i wish i had the knowledge to do AS/400 programming, there's an opening in town right now
-
wyatt8740
would be good money for a while
-
wyatt8740
hmm, nspr is throwing up errors from 'as' while building. how do i get more details (like the invocation?)
-
frg_Away
wyatt8740 Did mostly OS/390 z/OS. PL/I TSO IMS DB2 CTRL-M and TELON generator. I still love PL/I
-
frg_Away
mach build -v ?
-
wyatt8740
ah, d'oh
-
wyatt8740
still used to make -f client.mk
-
wyatt8740
haha, that's funny
-
wyatt8740
i think this is happening because it's trying to use -fPIC
-
frg_Away
Needs NASM and still YASM. Might need to convice Bill to port the patches needed to kick out yasm :)
-
wyatt8740
wasn't nasm x86-only? or am i forgetting everything again
-
wyatt8740
I only really do assembly on the 68k
-
frg_Away
Don't know. Will find out if I find the time for doing patches to allow for mac or Windows arm builds. Probably mac first.
-
wyatt8740
hmmm. i have no idea where it's even getting -fPIC from
-
wyatt8740
not entirely unheard of apparently
mozilla/nixpkgs-mozilla #278
-
wyatt8740
too bad the explanation doesn't make sense to me
-
wyatt8740
ah, thanks grep. `old-configure: ASFLAGS="$ASFLAGS -fPIC"`
-
wyatt8740
` 0:01.40 Fatal error: invalid listing option `,'`
-
wyatt8740
still got work to do i guess
-
buc
frg: sent you an email with better fix.
-
buc
A time to open a bug I believe :)