good morning WG9s

good morning!

all I need is to know what you want me to make this Buoy do postimg.cc/crdZQsDb

nsITobin: if you are asking me please send this in other than an image. It is completely illegible to me

can you click on it to zoom it out?

or is postimage just crap

here is imgbb ibb.co/JxXSCGT

its just a window with a menubar a browser control and devtools

i trid soming before i asked for other than image

the subwindow with the actual error is illegiable and zomming is of zero help

cantl you just give me the test of the error or a complete log from the build or something

I am confused.

I need to reboot back to windows

hi nsITobin

sup

just poking around about pages on this Buoy component i whipped up a little while back on central.. considering hooking up that overlay emulation and spawning navigator.xhtml

last I attempted it was around 122

0730|12:44:29 <+nsITobin> here is imgbb ibb.co/JxXSCGT <- i.ibb.co/CJxSWYy/image.png if it helps

its cool thanks njsg

but I don't think this was about an error, just showing it working with comm-central?

ye .. i need to work on being more clear

postimg.cc has a download image link that, at least with "open in browser" can be opened as an image

ibb.co has an image_src <link> that can be reached from the website navigation bar

(under "More")

postimg seems to go down a lot or has lately

I just need a paste/img bin

nsITobin: are you applying the gitlab seamonkey-central-patches patch queues?

yeah

that does a lot

OK just on first blush my take was maybe not appling the patches for bug TOP-1611647-1-rename-xul-suite.patch and TOP-1611647-2-fix-xul-references-suite.patch

shich have to do with changing comm-central to handle change .xul files to .xhtml

I think I can make suiteglue partly work by jsm'ing it but that is quick on the slate for removal and sys.mjs would ahve to swiftly follow to get seamonkey to start normally

least enough to spew errors

some of these issues are gonna still bare down with backports

but since errors seemded to be about .xhtml files i suspected this bug 2 parts one rename all files with .xul extensions to .xhtml the second hanges the filenmes in all the references

if you are adding more .xul files you might need my script that redoes the renames and rereferences .xul files do not work in central

the sehll script here takes new changes into account to fix the patch for this gitlab.com/frg/seamonkey-central-pa…scripts/xul2xhtml.sh?ref_type=heads

WG9s: I been studying very closely what they been up to dpaste.org/oGGiG

it irks me

but.. i learned

anyway if doing central patches don;t use .xul extenyion use .xhtml instead

indeed

still seems dumb

an internal extension shouldn't REALLY matter

just what a few lines to make sure it opens in the right parser

but .xhtml is a web standard thing .xul is a mozilla thing

yeah and about: protocol hides most chrome urls ;)

so the stupid scritp changes all .sul files inthe repo to .xhtml and a scond patch that finds all referwnces in the code to .sul files and changes thenm to .xhtml

still this is about as XUL as xhtml can get anymore as near as I can figure from other examples throughout the codebase

they have some xbl replacements as xul customelements but that's just transitional

you would think who cares but they removed the special code for .xul fiels form m-c becuase id you just rename them to .xhtml the code for that takes care of them

so no longer need special case for .xul files

sad

maybe they need to chage it to say there is only xytml

shtml

xhtml

third time is the charm

register the domain name redirect it back to XUL

;-)

WG9s: IF IT STOPPED HERE.. i MIGHT be willing to embrase it but we know it doesn't

that's the thing about modern mozilla

as a .. whatever it is

I am big on redirects and meta refresh pages in the archive for my nightly builds becuase seems no way to do symlinks

symlinks can be a pain

i think it is funny out of all of it.. composer of all things is SLIGHTLY functional

so like was about to send you an exapmle but seems the seamonkey project archive site is now down, or at least not reachable via version from the US

u'd think Mozilla would ressurect the editor ui for them selves for chrome html usage

so like archive.seamonkey-project.org/nightly/comm-253-latest/index.html is an index.html page with a meta refresh instead of archive.seamonkey-project.org/nightly/comm-253-latest being a symlink to the latest build

nsITobin: so already a typo in my automation scipt change

somehow pusddd instead of pushd

well pushdd instead of pushd but you get the idea

frg_Away: I am at the mangling my own patches stage of my hg mq learning

have mq fun

i am :)

aaand a pull from c-c just busted me

says run mach tb-rust vendor do it and it fails

0:00.65 [INFO] Updating gkrust in our workspace

error: invalid escape sequence

expected `b`, `f`, `n`, `r`, `t`, `u`, `U`, `\`, `"`

--> Cargo.toml:23:23

|

23 | cmake = { path = "..\..\build\rust\cmake" }

GREAT system..

seems frg broke something whn yu ceck-out c-c dot from changeset 30c2520e5b0ec3fe26c2c8c69f24352732f37463

foit

that is the last changeset from before when frg screwed the pooch

not sure his stuff is the reason but looks suspicious

how does adding a few urls to debugqa do this?!

not sure at all

looks like rust

just bustage seems c-c only and only these 3 changes

now i get a toml.decoder.TomlDecodeError: Reserved escape sequence used (line 18 column 1 char 294)

checking out the eatlier changeset did not help so must be a m-c change that broke something

I suspect something was chacked in no m-c that broke enable thunderbird-rust

try building with that disabled maybe?

oh stuipd me tested earlier changes wrong

testign correctly based on rgular chanset in about 5 mins

but then about 20 mins later to know

oh well it IS just central anyway.. I don't need to get all I MUST DO THIS NAOW about it .. learned some more progressed fullilling one of the things I said i would do.. learned more about hg mq.. all in all a good day of hackin on codestuffs

JUt seems to be boken when i ws trying to test tomething

if it is the frg stuff i would think mostlikely Bug 1909853 becuase it touches mailnews which is where the enable-thunderbird-rust comes into play

but was there a change in Cargo.toml?

yes

and a new rust component added

in cc

hold on i may be dumb and thowing WG9s off.. always a possibility

but m builds are failing and so are the thunderbird builds failurews started with frg's 3 patches landed

so unless something landed around same time on m-c that borke this is supect the feg patches

frg

nsITobin: frg is gone for the day i am trying to taise IanN_Away

well i was on rev 42116 i see another sync vendoreed rust libs with moziilla central, repackage thunderbird as a debian package, my ui patch, a chat patch and esr links to debugqa.. how can frg's patches break hg like is it something in the header? or

I am gonna try purging .mozillabuild .cargo and .rustup see if completely "fresh" it works

well could caus the error i am seeing not sure why your weeoe looks differnet

wait, why the backslashes?

I am on windows

atm

but

aren't all paths in mozilla regular slashes

to the build system

save internal python

maybe that's why

no if you are in windows seems either slash works sometimes

this is one of the things i complained about in the dos days

Rust dependencies are out of sync. Run `mach tb-rust vendor`

Microsoft decided to go the DWC VMS ay of things fielnames are case insensitive options use a / as the dleimiter so not avilable to use as a path delimeter

DOS itself could handle another separator, third party software is probablya trickier quesiton

run it

it fails

now Windows NT, my guess would be it translates internally too?

dos version one allowed you to change the command option delimeter and i ran at the time changing it to - but in vrsion 2 they elimiated that and came up with path delimiter is \

and i don;t remeber there being a way to change

they seems to be tageting making it be a DEC VMS replacement on a smaller platform

all worked berrer in the msdos they stole to sell to IBM

tomake pcdos

nsITobin: is there more than one error about escape sequences? because Cargo.toml:23:23 looks a bit off

I mean, the code printed afterwards has a *valid* escape at column 23

well my build output is a bit manged because i am not using the msys2 terminal but just through oldstyle conhost..

njsg WG9s here dpaste.org/eRnpw/raw

cleaned it up.. slightly

this is tip *-central + patches

now

ok that is now the same error i get has to do with tying to get the audioipc2-client rust package s i think is thunderbird-rust related

well good that's progress we both see the same thing.. that's a good sign.. for sanity at least

as you know I been encountering pretty much every quirk there is that can block someone building

nsITobin: are you ring to build seamonkey?

atm yes

could try Mozilla Thunderthighs

i mean

Thunderbird

does it word if you disable-thunderbird-rust?

well won't opengpg fail to build without thunderbird rust?

pgp

whatever it is

i am trying to build under linux 64 so fails if disalbe-thunderbitd-rust anyway

no it was that exchange protocol

lst i kne seamnkey did not try to build openpgp becuase under 2.53 it can still use the enigmail add-on

one of the things needed to change for central it needs to build openpgp i think

but it currently is not so pgp for seamonkey central does not do pgp as far as i know

becuase it teuite an add-on of a type no longer supported and does not include the opengpg that TB does

they did remove a dependancy on the appcore component part that shouldn't have been linked from mailnews

that is what I thought the current state of this was

also configure completed successfully

which means m-c is in .. well mozorder

and thunderbird busted comm

one of the three bugs frg landed at the same time the TB build issue began removed a lot of supposeldy unused mailnews code i am thinking pethaps he got a bit carried away in removal

ibf my build in progress sorrks then when you clone form comm-central speciy changeset

30c2520e5b0ec3fe26c2c8c69f24352732f37463

that is last good build before the bustage if you look at treeherder for comm-cdntal

I do not buy it's frg's changes unless the patches them selves managled comm-central the repo somehow.. his changes don't affect anyone or rust

my build still fails so wonder if something changed onm-c that needs a redo syncing

I am gonna try a local backout of the last thunderbird rust-sync

see what that does

oh look configure completed

WG9s:

now doesn't mean it won't compile error out.. but

i would suspect one of these patches as causing this

no still saying crap about audioipc2-client

unable to update

and i bet this one since it way audio ipc right in the chck-in comment

0:10.83 Unable to update github.com/mozilla/audioipc?rev=c20…67b875b46ce7c395ca172fae5b#c200b1ef

it would seem this reuites a comm rust sync which was not performed

why does that url have an anchor on it?

this? mozilla/audioipc c200b1e.diff

nsITobin: the anchor is the start of the commit hash. I chalk it down to github being github

why is the address not /commit/? or is that being output by the build system?

by the build system

couldn't update

yeah i don't understand how rust crates work

nsITobin: so if you backout the patch for bug 1910523 that should fix it untll we can get a new sync run it is not an extra sync that is the issue is a missing one that bug requires a sync which never happened.

wasiest way to fix for now is to backit out

easiest

nsITobin: TB folk figured it out and just did a sync so all should be well now

oh

i was just about to hopefully confirm it was all their fault

:P

I had to revert the mozilla part and the comm part to return cargo back to its prior state.. guess what I get to do now WG9s lol

put it all back?

YES!

here we go WG9s web.archive.org/web/20130116073227/…a.org/showbuilds.cgi?tree=SeaMonkey even has burning builds