<farkuhar>
According to the tint2 manpage, the first time you run the program it will create the file $HOME/.config/tint2/tint2rc (unless already present). But on two systems that I tested, the file that actually gets created is $HOME/.config/tint2/t/nt2rc. I think some of the #defines around line 1370 of tint2-17.1.3/src/config.c are going haywire.
<farkuhar>
Would one of the OpenBSD users do me the favor of testing tint2 on their machines? It would help to identify whether the bug is actually in the tint2 source, and not in the C libraries or the compiler toolchain.
<SiFuh>
ls .config/tint2/tint2rc
<SiFuh>
.config/tint2/tint2rc
<SiFuh>
tint2: could not find a config file! Creating a default one.
<farkuhar>
SiFuh: thanks for testing. Our CRUX port downloads from a forked repo that has been bumped to version 17.0.3, so maybe the bug got introduced by the new upstream developer.
<farkuhar>
17.1.3, in fact.
<SiFuh>
I don't use task bars
<SiFuh>
I have one in fluxbox but it just shows me what desktop I am on and the time
<farkuhar>
diff between tint2-17.{0.2,1.3}/src/config.c shows that the new upstream developer got rid of several g_build_filename calls in favor of a PATH_TEMPLATE macro. The latter implementation seems to be broken.
<farkuhar>
Old version: path1 = g_build_filename(g_get_user_config_dir(), "tint2", "tint2rc", NULL);
<farkuhar>
New version: #define PATH_TEMPLATE ("%s" G_DIR_SEPARATOR_S "tint2" G_DIR_SEPARATOR_S "tint2rc")
<SiFuh>
Backdate it, if it works, make a patch. Then push to to jmq
<farkuhar>
It's in the contrib repo, actually.
<SiFuh>
Isn't it a jmq port?
<SiFuh>
I am sure I saw John's name in the Maintainer line
<SiFuh>
└─> grep John /usr/ports/contrib/tint2/Pkgfile
<SiFuh>
# Maintainer: John McQuah, jmcquah at disroot dot org
<farkuhar>
contrib is a repo where the maintainers tend to identify themselves. In the official repos, they can get away with the anonymity of "CRUX Xorg team" or "CRUX System team, core-ports at crux dot nu".
<SiFuh>
Yeah, but you don't fuck with someone elses port. So you can inform jmq of the issue and provide a patch if you find a way to better it.
<SiFuh>
beerman exploded when stenur edited his port in contrib
<farkuhar>
Am I being whooshed here? SiFuh is talking to jmq and he doesn't appear to know it. That would be like me telling SiFuh "don't mess with yenjie's ports, just submit a PR once you figure out how to fix them."
<SiFuh>
farkuhar: are you John?
<farkuhar>
yes
<SiFuh>
Hmm
<SiFuh>
Well change your name before you fuck with John's ports ;-)
<jmq>
The commit message says the intent was to "optimize strings allocation". So if I apply the reverse patch, should I expect the strings allocation to degrade? I'm not familiar enough with the internals of g_build_filename to have an informed opinion right now.
<SiFuh>
jmq: Actually, it seems that it can find the config and all is good but if there is no config it creates that one
<SiFuh>
Looks like XDG introduction lead to it
<jmq>
SiFuh: if there is no config, the file that gets created is $HOME/.config/tint2/t/nt2rc (at least on two systems that I tested). The intended target is $HOME/.config/tint2/tint2rc, and I'm wondering how the final i got replaced by a slash.
<SiFuh>
As I said. Seems it doesn't create it when tint2rc exists. It finds it. But if it doesn't exist it creates that bullshit one
<jmq>
Hehe, it's considered a bug if the program's behavior differs from what's described in the man-page. If the intended behavior is really to create the bullshit file t/nt2rc, then maybe I should patch the man-page instead. Somehow I doubt they intended to save in that bullshit location, though (because the resulting file is not even loaded at the next startup).
<SiFuh>
I reckon it is a bug
<jmq>
What is this _mkdir_if_missing(userpath) at line 1423 of the commit 5d784c24? userpath in the old code was supposed to terminate in "tint2/tint2rc", not to be a directory that needs creating.
<jmq>
The new code seems to assign the same meaning to userpath, so I can't see why they thought it was a valid argument for _mkdir_if_missing. None of the other added lines have side-effects that would truncate userpath to just the directory.
<SiFuh>
Satellite phones and packages are expensive
<zorz>
hahhaaha
<zorz>
what about elons sats?
<SiFuh>
zorz: I think you don't know how they work
<zorz>
no idea
<SiFuh>
:-)
<zorz>
how he calls them satellite X, space X ?
<zorz>
i know he is a provider of satellite communicatins
<SiFuh>
Thuray was good because they were geosynchronis which can be good depending on your location. Iridium's aren't and they have heaps more satellites. Unfortunately they are a US company.
<SiFuh>
Inmarsat are British. Many of theirs are geosynchronis
<SiFuh>
I preferred Thuraya because it was under Dubai
<SiFuh>
But they only had 3 satellites
<zorz>
inmarsat, i heard before
<SiFuh>
Good hearing
<SiFuh>
They are pretty silent up there in space
<zorz>
blowfish encryption send it with email... still good.
<zorz>
but real time communication..... everybody listens
<zorz>
fuck i just missed it cause i was open shorting solana.
<zorz>
i know it 72 hours now.
<zorz>
SiFuh: as you can understand i started using again my feed in twitter :) preparing a python script to have the feed in terminal so i do not have to open firefox.
<SiFuh>
Why would I understand
<SiFuh>
I don't even have nor have I ever had a twit account
<SiFuh>
I use Telegram, IRC, email, SMS and phone calls.
<zorz>
i only use it for crypto, on some specific accounts.
<zorz>
anyway.... till end of augost.... it will be steady towards downhill... getting ready for mid september.