<ppetrov^>
farkuhar, why don't we have depends=(...) in the Pkgfile but "Depends on:"
<ppetrov^>
have you found a discussion about it?
<cruxbot>
[core/3.7]: python3: updated to version 3.10.15
lavaball has quit [Remote host closed the connection]
<farkuhar>
ppetrov^: click on "Back to the thread" from the link I posted above, and you'll see one such discussion.
<ppetrov^>
thanks
<ppetrov^>
btw, will python2 be removed freom crux 3.8?
<ppetrov^>
if so, then all python3-* ports should become just python-*?
uwumeowmeownyaa has quit [Ping timeout: 260 seconds]
<ppetrov^>
pfff,... i have a bunch of python ports to create and really miss a tool 'pip2crux'
serpente has joined #crux
<cruxbot>
[opt/3.7]: stunnel: updated to version 5.73
<cruxbot>
[opt/3.7]: iwd: updated to version 2.21
<cruxbot>
[opt/3.7]: mod_php: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-zip: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-sqlite3: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-sockets: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-mysql: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-mbstring: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-gettext: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-gd: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-ftp: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-fpm: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php-fcgi: updated to version 8.3.11
<cruxbot>
[opt/3.7]: php: updated to version 8.3.11
<cruxbot>
[xorg/3.7]: mesa: updated to version 24.2.2
lavaball has joined #crux
<farkuhar>
Looking beyond the harsh opening lines of Anton's August 2006 reply, you can see him actually address the underlying issue: customizing ports. stenur's Gitea proposal #4 tries to defer the customization until running pkgmk, but Anton (anticipating deep42thought) binds it to `ports -u`: https://dpaste.com/CSRWUMU6S
<farkuhar>
Sorry for not linking to a more permanent URL. As I mentioned yesterday, the transition from pipermail to hyperkitty seems to have incurred some data loss, so I had to recover the full thread from my personal backups.
<ppetrov^>
"ghastly to death"
groovy2shoes has quit [Remote host closed the connection]
groovy2shoes has joined #crux
<SiFuh>
ppetrov^: Anton is zorzing ;-)
<farkuhar>
Anton and deep42thought both preferred to bind user customizations to the `ports -u` step, unlike stenur who wants the changes incorporated into pkgmk. But proposed patches to pkgmk have a long history of getting ignored. Consider Per's own suggestion to endow pkgmk with dependency awareness, or Vitaly's idea of a --dumpconfig flag.
<farkuhar>
More recently, FS#1852 brought up the problem that Pkgfile is sourced before signatures are verified, potentially allowing the check_signature() function to be overwritten by a malicious port. This issue remained at the top of the Flyspray landing page for years, before Flyspray was quietly retired in favor of Gitea.
<SiFuh>
farkuhar: Could always do it like they do with driver engines for downloading. Chuck an /etc/ports/plugins directory in that pkgmk could scan.
<SiFuh>
Wouldn't need much tweaking.
<farkuhar>
SiFuh: that's more or less what stenur is saying in Gitea #4.
<SiFuh>
Makes more sense, because then it would be up the user to decide if they want to utilize the new features or make their own. Could have a contrib/plugins section.
<SiFuh>
user-contrib/pkgmk-plugins
<farkuhar>
It's understandable that Per in 2004 might have expected more enthusiasm for incorporating new features into pkgmk. Back then we hadn't yet witnessed the languishing of the rich discussion that culminated in https://crux.nu/Wiki/PkgutilsIdeas without leading to much actual implementation.
<farkuhar>
Anton in 2006, aware of the ongoing PkgutilsIdeas discussion, might have been wary of proposing a change in pkgmk, and hence attached his suggestion to the `ports -u` step. Interestingly, around the same time as the PkgutilsIdeas discussion, complaints had surfaced about duplicate functionality (listing and diff) where the prt-get output deviated from the /usr/bin/ports output.
<farkuhar>
So if /usr/bin/ports was going to see its functionality slimmed down, eliminating listing and diff to focus on the networking part, Anton had reason to believe that `ports -u` was the natural place for incorporating local patches.
<farkuhar>
Heh, jw's reply to Per in 2004 deserves quoting: "I might not be the most neutral person here, but I kind of liked the different levels pkgmk and prt-get were working. ... I kinda like pkgmk's simplicity to not look beyond a port directory."
_0bitcount has joined #crux
<SiFuh>
I'd agree
serpente has quit [Remote host closed the connection]
_moth_ has quit [Ping timeout: 248 seconds]
groovy2shoes has quit [Remote host closed the connection]
groovy2shoes has joined #crux
ivandi has quit [Quit: WeeChat 4.4.1]
ivandi has joined #crux
<jaeger>
crux.ninja's web server is down for a bit of maintenance