From webdev at snerpton.net Wed May 6 09:10:04 2026 From: webdev at snerpton.net (Chris Ashton (webdev)) Date: Wed, 6 May 2026 10:10:04 +0100 Subject: =?utf-8?Q?Review_request_=E2=80=93_cloudflare-speed-cli=3A_new_po?= =?utf-8?Q?rt_=28v0=2E6=2E6=29_=2332587?= Message-ID: Hi, Can I request a review of my pull request for a new port submission: https://github.com/macports/macports-ports/pull/32587 This is my first submission, so I might not have followed the correct procedure! Cheers, Chris Ashton -------------- next part -------------- An HTML attachment was scrubbed... URL: From ferdy at ferdy.it Wed May 6 17:13:34 2026 From: ferdy at ferdy.it (Giuseppe Miceli) Date: Wed, 6 May 2026 19:13:34 +0200 Subject: selfupdate always updates macports bases even if the installed version is the latest Message-ID: <18D167EA-F778-4C32-9519-2899EDD6CBBE@ferdy.it> ciao, on a specific machine, everytime i selfupdate macports, it detects correctly the installed macports base versions installed and available and even if they are the same version it updated macports base (see iterm dump below pasted). don?t know if it?s my fault or some stale file around or else? am i supposed to open a ticket on trac or is it a trivial thing? thank you very much in advance, ? ferdy at kamikaze:~$ sudo port -svpf selfupdate && port outdated Password: ---> Checking for newer releases of MacPorts % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 63 100 63 0 0 328 0 --:--:-- --:--:-- --:--:-- 329 MacPorts base version 2.12.5 installed, MacPorts base version 2.12.5 available. ---> MacPorts base is outdated, installing new version 2.12.5 ---> Attempting to fetch MacPorts 2.12.5 source code from https://github.com/macports/macports-base/releases/download/v2.12.5/MacPorts-2.12.5.tar.bz2 % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0 100 24.9M 100 24.9M 0 0 9426k 0 0:00:02 0:00:02 --:--:-- 11.0M Attempting to fetch signature from https://github.com/macports/macports-base/releases/download/v2.12.5/MacPorts-2.12.5.tar.bz2.sig % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0 100 155 100 155 0 0 456 0 --:--:-- --:--:-- --:--:-- 456 ---> Extracting MacPorts 2.12.5 ---> Installing new MacPorts release in /opt/local as root:admin; permissions 0755 From dluke at geeklair.net Wed May 6 18:03:17 2026 From: dluke at geeklair.net (Daniel J. Luke) Date: Wed, 6 May 2026 14:03:17 -0400 Subject: Broken cpp? (apple clang 17) Message-ID: <07169547-3945-407C-8415-C1A046C66526@geeklair.net> People may already know this, but it looks to me like `clang -E` (the c-preprocessor) Apple's version 17.0.0 is somewhat broken. I noticed when trying to build apache httpd 2.4.67 on a macOS 15 host, that the generated server/exports.c appeared to be truncated at newlines from AP_DECLARE_NONSTD macros (from include/http_config.h). Building with macpots clang-22 works fine, though (CC=/opt/local/bin/clang-mp-22 CPP=/opt/local/bin/clang-cpp-mp-22). Anyone else seeing similar? -- Daniel J. Luke From lists at schamschula.com Wed May 6 18:32:29 2026 From: lists at schamschula.com (Marius Schamschula) Date: Wed, 6 May 2026 13:32:29 -0500 Subject: Broken cpp? (apple clang 17) In-Reply-To: <07169547-3945-407C-8415-C1A046C66526@geeklair.net> References: <07169547-3945-407C-8415-C1A046C66526@geeklair.net> Message-ID: <738E5E7E-1A22-457C-83A8-CEB40175DF7B@schamschula.com> Nope, I didn?t see this issue when I first build it on my Sequoia ARM machine, before committing the update to GitHub. Speaking of the pre-built binaries: several build bots (including macOS 15 x86_64) claim a successful build, but didn?t actually build anything. See: https://build.macports.org/builders/ports-15_x86_64-watcher/builds/18919/steps/subports/logs/stdio Marius -- Marius Schamschula > On May 6, 2026, at 1:03?PM, Daniel J. Luke wrote: > > People may already know this, but it looks to me like `clang -E` (the c-preprocessor) Apple's version 17.0.0 is somewhat broken. > > I noticed when trying to build apache httpd 2.4.67 on a macOS 15 host, that the generated server/exports.c appeared to be truncated at newlines from AP_DECLARE_NONSTD macros (from include/http_config.h). Building with macpots clang-22 works fine, though (CC=/opt/local/bin/clang-mp-22 CPP=/opt/local/bin/clang-cpp-mp-22). > > Anyone else seeing similar? > -- > Daniel J. Luke > Marius -- Marius Schamschula -------------- next part -------------- An HTML attachment was scrubbed... URL: From fw at fwright.net Wed May 6 19:56:02 2026 From: fw at fwright.net (Fred Wright) Date: Wed, 6 May 2026 12:56:02 -0700 (PDT) Subject: selfupdate always updates macports bases even if the installed version is the latest In-Reply-To: <18D167EA-F778-4C32-9519-2899EDD6CBBE@ferdy.it> References: <18D167EA-F778-4C32-9519-2899EDD6CBBE@ferdy.it> Message-ID: On Wed, 6 May 2026, Giuseppe Miceli wrote: > on a specific machine, everytime i selfupdate macports, it detects correctly the installed macports base versions installed and available and even if they are the same version it updated macports base (see iterm dump below pasted). > > don?t know if it?s my fault or some stale file around or else? am i supposed to open a ticket on trac or is it a trivial thing? > > thank you very much in advance, > ? > > ferdy at kamikaze:~$ sudo port -svpf selfupdate && port outdated [...] It's doing exactly what you told it to do with -f. Also note: -s is a NOP on selfupdate, since it (unfortunately) always builds from source. -p is probably a NOP, since AFAIK, selfupdate is treated as a single operation. Fred Wright From ryandesign at macports.org Wed May 6 23:10:37 2026 From: ryandesign at macports.org (Ryan Carsten Schmidt) Date: Wed, 6 May 2026 18:10:37 -0500 Subject: Broken cpp? (apple clang 17) In-Reply-To: <738E5E7E-1A22-457C-83A8-CEB40175DF7B@schamschula.com> References: <738E5E7E-1A22-457C-83A8-CEB40175DF7B@schamschula.com> Message-ID: On May 6, 2026, at 13:33, Marius Schamschula wrote: > > Speaking of the pre-built binaries: several build bots (including macOS 15 x86_64) claim a successful build, but didn?t actually build anything. Because they had already built it before the web hooks could be delivered, because I scheduled builds manually, because of https://trac.macports.org/ticket/73701#comment:23 > See: https://build.macports.org/builders/ports-15_x86_64-watcher/builds/18919/steps/subports/logs/stdio "Excluding apache2 because it has already been built and uploaded to the public server" -------------- next part -------------- An HTML attachment was scrubbed... URL: From ryandesign at macports.org Wed May 6 23:36:41 2026 From: ryandesign at macports.org (Ryan Carsten Schmidt) Date: Wed, 6 May 2026 18:36:41 -0500 Subject: selfupdate always updates macports bases even if the installed version is the latest In-Reply-To: References: Message-ID: On May 6, 2026, at 14:56, Fred Wright wrote: > On Wed, 6 May 2026, Giuseppe Miceli wrote: > >> ferdy at kamikaze:~$ sudo port -svpf selfupdate && port outdated > [...] > > It's doing exactly what you told it to do with -f. > > Also note: > > -s is a NOP on selfupdate, since it (unfortunately) always builds from source. > > -p is probably a NOP, since AFAIK, selfupdate is treated as a single operation. Although Fred is right that -s and -p do nothing for selfupdate and -f intentionally causes the symptom you're reporting, it should also be noted that you're not expected to need to use most of these flags regularly or at all when installing or upgrading ports either and some of them may be harmful. The -p flag proceeds past a build failure and continues to build other ports, even if they depend on the failing port. If so, and if you already have an older version of the failing dependency installed, it'll use that. Depending on why the dependency was updated, this may result in the other port build failing (or worse, building without error but being subtly incorrect). You should only use this flag if you are certain that the changes in the dependency are not significant to the other port. This is not information we expect most users to know since it could take a lot of investigation to find out, so most users should not use this option. The -f flag forces overwriting files that already existed that MacPorts didn't know about, but that shouldn't happen, and if it does you'd want to investigate why those files were there before blasting past it. The -s flag forces a port to build from source rather than download a binary. Ordinarily you would only do this if you were diagnosing a port problem. If you always want to build ports from source on principle, you can set "buildfromsource always" in macports.conf. The -v flag produces verbose output and this (or the -d flag for even more debug output) are fine to use if you want that much info in your terminal. You can also see this information later in the port's log file, unless the build was successful and MacPorts was configured (as it is by default) to delete logs on successful builds. From fw at fwright.net Wed May 6 23:43:22 2026 From: fw at fwright.net (Fred Wright) Date: Wed, 6 May 2026 16:43:22 -0700 (PDT) Subject: selfupdate always updates macports bases even if the installed version is the latest In-Reply-To: References: Message-ID: On Wed, 6 May 2026, Ryan Carsten Schmidt wrote: > The -p flag proceeds past a build failure and continues to build other > ports, even if they depend on the failing port. If so, and if you Which of course is only true because -p erroneously implies -n, instead of behaving like make's -k, which keeps single errors from blowing the whole run without ignoring dependencies. If you want the current -p behavior, you should have to specify both -p and -n. Fred Wright From ryandesign at macports.org Wed May 6 23:46:54 2026 From: ryandesign at macports.org (Ryan Carsten Schmidt) Date: Wed, 6 May 2026 18:46:54 -0500 Subject: selfupdate always updates macports bases even if the installed version is the latest In-Reply-To: References: Message-ID: On May 6, 2026, at 18:43, Fred Wright wrote: > >> On Wed, 6 May 2026, Ryan Carsten Schmidt wrote: >> >> The -p flag proceeds past a build failure and continues to build other ports, even if they depend on the failing port. If so, and if you > > Which of course is only true because -p erroneously implies -n, instead of behaving like make's -k, which keeps single errors from blowing the whole run without ignoring dependencies. If you want the current -p behavior, you should have to specify both -p and -n. That seems like it would be a reasonable change... unless I'm forgetting something. From dluke at geeklair.net Thu May 7 12:22:13 2026 From: dluke at geeklair.net (Daniel J. Luke) Date: Thu, 7 May 2026 08:22:13 -0400 Subject: Broken cpp? (apple clang 17) In-Reply-To: References: Message-ID: I should clarify that I see this on an Intel Mac. I guess I can see about reinstalling Xcode and command line tools on it if no one else is seeing this behavior (I haven?t seen it building other software). -- Daniel J. Luke > On May 6, 2026, at 7:10?PM, Ryan Carsten Schmidt wrote: > > ? >> On May 6, 2026, at 13:33, Marius Schamschula wrote: >> >> Speaking of the pre-built binaries: several build bots (including macOS 15 x86_64) claim a successful build, but didn?t actually build anything. > > Because they had already built it before the web hooks could be delivered, because I scheduled builds manually, because of https://trac.macports.org/ticket/73701#comment:23 > > >> See: https://build.macports.org/builders/ports-15_x86_64-watcher/builds/18919/steps/subports/logs/stdio > > "Excluding apache2 because it has already been built and uploaded to the public server" -------------- next part -------------- An HTML attachment was scrubbed... URL: From dluke at geeklair.net Thu May 7 21:10:10 2026 From: dluke at geeklair.net (Daniel J. Luke) Date: Thu, 7 May 2026 17:10:10 -0400 Subject: Broken cpp? (apple clang 17) Message-ID: <8379BE70-C9FB-446B-895E-5A7EC2384549@geeklair.net> ?As a follow-up, reinstalling Xcode and the command line tools did fix this. I?m not sure how this box ended up in a bad state - but hopefully this helps others if they see the same issue. Thanks everyone - as usual the community here is most helpful :) -- Daniel J. Luke -- Daniel J. Luke > On May 7, 2026, at 8:22?AM, Daniel J. Luke wrote: > ?I should clarify that I see this on an Intel Mac. > > I guess I can see about reinstalling Xcode and command line tools on it if no one else is seeing this behavior (I haven?t seen it building other software). > -- > Daniel J. Luke > >> On May 6, 2026, at 7:10?PM, Ryan Carsten Schmidt wrote: >> ? >> On May 6, 2026, at 13:33, Marius Schamschula wrote: >>> Speaking of the pre-built binaries: several build bots (including macOS 15 x86_64) claim a successful build, but didn?t actually build anything. >> >> Because they had already built it before the web hooks could be delivered, because I scheduled builds manually, because of https://trac.macports.org/ticket/73701#comment:23 >> >> >>> See: https://build.macports.org/builders/ports-15_x86_64-watcher/builds/18919/steps/subports/logs/stdio >> >> "Excluding apache2 because it has already been built and uploaded to the public server" -------------- next part -------------- An HTML attachment was scrubbed... URL: From macintosh at mathiesen.info Sat May 9 13:14:12 2026 From: macintosh at mathiesen.info (Bjarne D Mathiesen) Date: Sat, 9 May 2026 15:14:12 +0200 Subject: php: PR32175, PR31982, PR31991 In-Reply-To: <875cd325-f4ae-4e0f-b282-c2e210b94d2c@mathiesen.info> References: <875cd325-f4ae-4e0f-b282-c2e210b94d2c@mathiesen.info> Message-ID: Latest notes: https://github.com/macports/macports-ports/pull/32175 php: update to @8.2.31, @8.3.31, @8.4.20, @8.5.5 autoconf @2.73 C23 fix +++++++ https://github.com/macports/macports-ports/pull/32645 php-sqlsrv: update to 5.13.1 Doesn't install on the GitHub Actions CI as arm64 isn't supported I've got build logs if you want to see them ++++++ https://github.com/macports/macports-ports/pull/32643 php-oracle: new Portfile @1.2.0 Doesn't install on the GitHub Actions CI due to the inablilty to install oracle-instantclient (https://trac.macports.org/ticket/73854) I've got build logs if you want to see them -- Bjarne D Mathiesen Kors?r ; Danmark ; Europa ----------------------------------------------------------------------- denne besked er skrevet i et totalt M$-frit milj? MacPro 2010 5.1 ; OpenCore + macOS 15.7.4 Sequoia 2 x 3,46 GHz 6-Core Intel Xeon ; 192 GB 1333 MHz DDR3 ECC RDIMM AMD Radeon Vega VII 16 GB From n_larsson at yahoo.com Sat May 9 14:50:55 2026 From: n_larsson at yahoo.com (Nicklas Larsson) Date: Sat, 9 May 2026 16:50:55 +0200 Subject: Request to rebuild py-pyqt5 References: <5713F134-4EF9-4EB2-8025-9DD93021259D.ref@yahoo.com> Message-ID: <5713F134-4EF9-4EB2-8025-9DD93021259D@yahoo.com> Hi, Could someone with access please start a rebuild of py{310-314}-pyqt5 on macOS 14, Sonoma, arm64 on the buildbot server? Thank you in advance, Nicklas