<farkuhar>
remiliascarlet: not a spelling error, just a choice to use one nick on IRC and another name for the published repo.
<farkuhar>
But sajcho only makes brief appearances on IRC, so maybe you weren't there to notice any recent drive-by postings.
<r0ni>
farkuhar: the prob with the pinentry port is the script included with it. I can't modify it but it lists a pinentry-gtk2 which doesn't exist when building it, the bin is pinentry-gnome3. I had originally created my own port fixing this as 'pinentry-gtk' but pkgadd refuses to install as it finds the same file in the 'pinentry' port unless -f is passed since i'm overwriting the pinentry port /usr/bin/pinentry. My solution as of now is to just
<r0ni>
completely overwrite the port that way it won't fail and contains a working script
<r0ni>
any user who uses pinentry and has xfce wouldn't be none the wiser as it includes the ncurses and gnome3 binary instead of just the ncurses. and seems to be the correct way to overlay ports anyway it appears
<r0ni>
basically, i'm doing what i'm doing now so a prt-get sysup works without manual intervention. which is important in my eyes
<r0ni>
a user not interested in the modified port wouldn't be using my repo anyway, sure there's ways I can make another different named pkg, but I can't fix the script in crux repo myself so my solution is to just overwrite it
<farkuhar>
Does the opt/pinentry script only need to be circumvented while building your port, or after installation as well? If it only matters at build time, you can just temporarily redefine PATH in your Pkgfile and put a modified script into a directory that appears earlier in PATH.
<r0ni>
the script is just a wrapper for calling pinentry
<r0ni>
it erronously has pinentry-gtk2 in it when the bin that is created is named pineentry-gnome3
<r0ni>
pinentry*
<r0ni>
which prevents it from working properly
<r0ni>
i could have users modify their paths or put the file elsewhere but that seems overkill when i can fix it myself and move on with my life OR better yet, whoevers port that is, could fix it and then i wouldn't have to circumvent anything
<r0ni>
anyway, I do apologize, but i've spent more than enough time on a simple port that could have been spent on other, more demanding ports, had the original in the crux repo not been gimped to begin with. so forgive my attitude. my solution is my solution and i don't think i'll be going out of my way to continue changing it. it works as-is and i'm contempt with that.
<farkuhar>
r0ni: thanks for pointing me in the right direction. I suggest you drop beerman a note over in #crux-devel and suggest that he incorporate your fixed wrapper into the official pinentry port. Nobody has touched the wrapper since _tek's update in 2016; beerman might appreciate the heads-up that it's out of date.
<farkuhar>
The Pkgfile format has a Maintainer line for exactly this reason. If you see something the matter with an official port, contact that person. Can you imagine the chaos it would cause to a CRUX user's filesystem, if repo maintainers just put arbitrary sed and tee commands into their {pre,post}-install scripts because they're too timid to propose the changes through official channels?
<remiliascarlet>
farkuhar: I'm not sure, but looking at the domain it's hosted on, and the maintainer name, I really doubt that he or she would even be here. It's a well known personality on the Japanese internet, and they only speak Japanese.
<farkuhar>
remiliascarlet: that reminds me, it was a Japanese developer who ushered into the world the text-mode browser w3m. Seeing that our contrib port of w3m is marked unmaintained, it would be fitting to see one of the Japanese repos adopt the w3m port if it eventually gets dropped from contrib.
<remiliascarlet>
That person introduced Lynx rather, but if it wasn't for him/her, I wouldn't even know about Crux.
<remiliascarlet>
Suwako made a blog post about how to make ports in Crux, which prompted me into looking it up, and I fell in love with the simplicity of it.
<sajcho>
farkuhar: good mesa build patch -https://dpaste.com/2CNDPWWDU , if you have time can you test it? Thank you.
<sajcho>
It is from alpine linux.
sajcho has quit [Quit: Client closed]
<r0ni>
i sent beerman an email, but beware, every mail i've sent for weeks has come back to me even tho my server passes all the email checks i've ran, and i'm tired of fighting with it, so luck will have it send
<SiFuh>
r0ni: beerman isn't real. He doesn't actually exist
<r0ni>
lol not to him, to anyone
<r0ni>
i get emails all day but anything i send fails some check somewhere and i've begun to just hate email
<SiFuh>
I think you may have a problem with your email client or server then
<r0ni>
it passes all the checks i run on it, and i'm honestly fine not sending emails lol
<SiFuh>
Better to have it set to not receive ;-)
<r0ni>
while actually true, you'd be surprised at how little spam i get, and the server been online for years
<r0ni>
and i'm not even running spamassasin
<SiFuh>
Only the emails I use for ebay/paypal and shopping have spam.
<r0ni>
ya only one address gets any at all, it's always crypto shit
<farkuhar>
r0ni: don't use email to contact beerman, then. Just hop over to #crux-devel and let him know your issue.
<SiFuh>
r0ni: don't listen to farkuhar. beerman is a bot
<SiFuh>
He's just offloading all work onto it
<farkuhar>
r0ni: why even bother maintaining your own email server? Let some other sysadmin deal with the obstacles that the FAANGs have put in the way of this federated communications system, and then use a non-interactive SSH session to initiate sendmail on one of those machines whenever you have something to send.
<farkuhar>
sajcho's patch for xorg/mesa looks good. In the new footprint, /usr/lib/dri/ will contain mostly symlinks, rather than all the hardlinks that our mesa port currently creates.
<SiFuh>
Yeah, farkuhar is right. Download WIndows and register with hotmail.
<SiFuh>
I run my own mailserver
<farkuhar>
I'd like to double-check my reading of what teodor is asking for in FS#1843. A unified listing of all ports not found in the active repos AND all ports not compatible with the requested action, no matter the number of ports passed as args (one vs. many) or whether dependency resolution was requested (depinst vs. install)?
<SiFuh>
That sounds like a job for debian. 'AND all ports not compatible with the requested action'
<farkuhar>
It's easier to address the divergent behaviour of prt-get --test in the case "one port vs. many ports". I also think I've narrowed down the cause of divergent behaviour in depinst vs. install: there was an early exit that prevented executeTransaction() from happening when dependencies could not be satisfied by the active repos.
<farkuhar>
teodor's feature request is much easier to incorporate into my Perl-based rewrite. In the C++ codebase I constantly have to reorient myself and remember which data structures are available in each method, and which side-effects of a subroutine are visible to the parent.
<r0ni>
farkuhar: because we have to take the power back, and until recently i've loved running my own email server, only something is bugged and i haven't found it yet (i also haven't really looked) cuz i'm busy with a million other things and i rarely send emails
ukky has joined #crux
SiFuh has quit [Remote host closed the connection]
SiFuh has joined #crux
<ukky>
If you have local ip routing tables, make sure you re-apply them at a new table location after iproute2 update. Location has changed from /etc/iproute2/rt_tables to /usr/lib/iproute2/rt_tables
SiFuh has quit [Quit: leaving]
lavaball has quit [Remote host closed the connection]