From dev at sethm.id.au Tue Sep 1 06:46:34 2026 From: dev at sethm.id.au (Seth McDonald) Date: Tue, 1 Sep 2026 16:46:34 +1000 Subject: Managing multiple patches for a source file Message-ID: <3CA62258-19B5-496D-B812-B0FD555AE866@sethm.id.au> Hi all, Following the MacPorts guide on patching source code (see [0], section 4.6.2), it encourages using separate patchfiles for each logical change and says to diff against the original file. Given this, I'd like to ask what the about the situation in which multiple logical changes need to be applied to the same source file. Since each change may alter the line count of the file, I worry that just diffing each change against the original file may cause later patches to not apply due to the expected line numbers being changed. I can think of three ways this can be managed: (0) Keep all changes to the same source file within the same patchfile. This way the problem never occurs, but removes the benefit of logically separate patches. (1) Diff each change against the original file. Perhaps I'm wrong and MacPorts is actually smart enough to handle sequential patches which each alter line numbers without the patchfiles accounting for it. (2) Diff each change against the previous change. This defines an order in which the patchfiles must be applied. This option makes the most sense in my head. Is one of these three methods the "correct" way in which multiple patches should be managed? Take care, Seth McDonald. [0] -- E9D1 26A5 F0D4 9DF7 792B C2E2 B4BF 4530 D39B 2D51 From jmr at macports.org Tue Sep 1 09:27:27 2026 From: jmr at macports.org (Joshua Root) Date: Tue, 1 Sep 2026 19:27:27 +1000 Subject: Managing multiple patches for a source file In-Reply-To: <3CA62258-19B5-496D-B812-B0FD555AE866@sethm.id.au> References: <3CA62258-19B5-496D-B812-B0FD555AE866@sethm.id.au> Message-ID: On 1/9/2026 16:46, Seth McDonald via macports-dev wrote: > Hi all, > > Following the MacPorts guide on patching source code (see [0], section > 4.6.2), it encourages using separate patchfiles for each logical change > and says to diff against the original file. > > Given this, I'd like to ask what the about the situation in which > multiple logical changes need to be applied to the same source file. > Since each change may alter the line count of the file, I worry that > just diffing each change against the original file may cause later > patches to not apply due to the expected line numbers being changed. > > I can think of three ways this can be managed: > > (0) Keep all changes to the same source file within the same patchfile. > This way the problem never occurs, but removes the benefit of logically > separate patches. > > (1) Diff each change against the original file. Perhaps I'm wrong and > MacPorts is actually smart enough to handle sequential patches which > each alter line numbers without the patchfiles accounting for it. The patch command is smart enough to apply patches at an offset from the original line numbers if lines have been added or removed. You will get patch failures if multiple patches are modifying the same lines, or even lines that are too close to each other. Applying with an offset is almost always fine, but patches that need a nonzero fuzz factor to apply should be avoided, as that has been known to result in code that compiles but has subtle bugs. > (2) Diff each change against the previous change. This defines an order > in which the patchfiles must be applied. This option makes the most > sense in my head. > > Is one of these three methods the "correct" way in which multiple > patches should be managed? All three options are acceptable as long as the patches apply cleanly. Use whichever one makes maintenance easiest in your case. If you end up managing patch series, you may find the quilt port useful. - Josh From dev at sethm.id.au Tue Sep 1 10:14:01 2026 From: dev at sethm.id.au (Seth McDonald) Date: Tue, 1 Sep 2026 20:14:01 +1000 Subject: Managing multiple patches for a source file In-Reply-To: References: <3CA62258-19B5-496D-B812-B0FD555AE866@sethm.id.au> Message-ID: <49F066E6-42CA-481B-8B75-13EF03269FC9@sethm.id.au> On Tue, 01 Sep 2026 at 19:27:27 +1000, Joshua Root wrote: > The patch command is smart enough to apply patches at an offset from the > original line numbers if lines have been added or removed. You will get > patch failures if multiple patches are modifying the same lines, or even > lines that are too close to each other. > > I see; thanks for the correction! > If you end up managing > patch series, you may find the quilt port useful. I'll look into it :) Take care, Seth McDonald. -- E9D1 26A5 F0D4 9DF7 792B C2E2 B4BF 4530 D39B 2D51 From l2dy at macports.org Tue Sep 1 14:58:37 2026 From: l2dy at macports.org (Zero King) Date: Tue, 1 Sep 2026 22:58:37 +0800 Subject: Sharing my ports tree development notes Message-ID: Hi, I'm sharing some notes I've written based on my experience contributing to the ports tree. Source code for the notes is at https://github.com/l2dy/ports-dev-notes and a generated PDF file is available on https://l2dy.github.io/ports-dev-notes/ for the time being. Regarding its license, while I intended to use the same license as https://github.com/macports/macports-guide, I didn't find a license file in it. Any suggestions? Feedback and discussions, either in reply to this email or through GitHub issues in my repo, are welcome! -- Best regards, Zero King From macintosh at mathiesen.info Tue Sep 1 19:26:14 2026 From: macintosh at mathiesen.info (Bjarne D Mathiesen) Date: Tue, 1 Sep 2026 21:26:14 +0200 Subject: outdated ports Message-ID: <85593f7c-62e6-4037-87ab-4760da74d332@mathiesen.info> Please see https://github.com/macports/macports-ports/pull/34418 -- Bjarne D Mathiesen Kors?r ; Danmark ; Europa ----------------------------------------------------------------------- denne besked er skrevet i et totalt M$-frit milj? MacPro 2010 5.1 ; OpenCore + macOS 15.7.8 Sequoia 2 x 3,46 GHz 6-Core Intel Xeon ; 192 GB 1333 MHz DDR3 ECC RDIMM AMD Radeon Vega VII 16 GB From mark at macports.org Tue Sep 1 19:40:27 2026 From: mark at macports.org (Mark E Anderson) Date: Tue, 01 Sep 2026 15:40:27 -0400 Subject: Sharing my ports tree development notes In-Reply-To: References: Message-ID: <647e7666-c078-4009-9795-defd3503ce70@app.fastmail.com> I?m interested in what the license is as well, I have an asciidoc version that I?m trying to update with how the portgroups work and bring up to date. Mostly for myself, since I?ve been trying to understand the internals better but never hurts to share. On Tue, Sep 1, 2026, at 10:58 AM, Zero King wrote: > Hi, > > I'm sharing some notes I've written based on my experience contributing > to the ports tree. Source code for the notes is at > https://github.com/l2dy/ports-dev-notes and a generated PDF file is > available on https://l2dy.github.io/ports-dev-notes/ for the time being. > > Regarding its license, while I intended to use the same license as > https://github.com/macports/macports-guide, I didn't find a license file > in it. Any suggestions? > > Feedback and discussions, either in reply to this email or through > GitHub issues in my repo, are welcome! > > -- > Best regards, > Zero King > ?Mark Macports Wiki Page | GitHub | Blog -------------- next part -------------- An HTML attachment was scrubbed... URL: From macintosh at mathiesen.info Tue Sep 1 19:45:39 2026 From: macintosh at mathiesen.info (Bjarne D Mathiesen) Date: Tue, 1 Sep 2026 21:45:39 +0200 Subject: [boost] v 1.89.0 1.90.0 1.91.0 1.92.0 In-Reply-To: References: Message-ID: <52dc1716-d917-4ed3-a09b-1407cfc202bb@mathiesen.info> I need some help / pointers I'm working on https://github.com/macports/macports-ports/pull/31402 but I'm stuck on a problem, where I don't know how to find the original cause of the problem. The problem is this : ################ DEBUG: system: cd "/opt/local/var/macports/build/boost189-0610ff64/work/boost_1_89_0" && /opt/local/var/macports/build/boost189-0610ff64/work/boost_1_89_0/b2 ... runtime-link=shared - j24 architecture=x86 ... notice: could not find main target j24 don't know how to make j24 ################ The problem is in the parameter specifing the # of CPU kernels In the build command it's '- j24', when it ought to be '-j24' It's working correctly in boost188, but in boost189..192 it's wrong The most curious thing is, that when I made the PR in march 2026 boost190 !was! working ; but now that one also fails :-( Which leads me to the possible conclusion, that something in MP itself has changed, that leads to this issue -- Bjarne D Mathiesen Kors?r ; Danmark ; Europa ----------------------------------------------------------------------- denne besked er skrevet i et totalt M$-frit milj? MacPro 2010 5.1 ; OpenCore + macOS 15.7.8 Sequoia 2 x 3,46 GHz 6-Core Intel Xeon ; 192 GB 1333 MHz DDR3 ECC RDIMM AMD Radeon Vega VII 16 GB From jmr at macports.org Tue Sep 1 21:06:34 2026 From: jmr at macports.org (Joshua Root) Date: Wed, 2 Sep 2026 07:06:34 +1000 Subject: [boost] v 1.89.0 1.90.0 1.91.0 1.92.0 In-Reply-To: <52dc1716-d917-4ed3-a09b-1407cfc202bb@mathiesen.info> References: <52dc1716-d917-4ed3-a09b-1407cfc202bb@mathiesen.info> Message-ID: <9bcd431b-a67a-44db-b01f-9f07bb8b1d97@macports.org> On 2/9/2026 05:45, Bjarne D Mathiesen via macports-dev wrote: > The problem is this : > ################ > DEBUG: system: > cd "/opt/local/var/macports/build/boost189-0610ff64/work/boost_1_89_0" > && /opt/local/var/macports/build/boost189-0610ff64/work/boost_1_89_0/b2 ... > runtime-link=shared - j24 architecture=x86 > ... > > notice: could not find main target j24 > don't know how to make j24 > ################ The extra white space is there in the build.args in your Portfiles. Line 224 in boost189 and the equivalent line in the others. - Josh From jmr at macports.org Tue Sep 1 21:20:04 2026 From: jmr at macports.org (Joshua Root) Date: Wed, 2 Sep 2026 07:20:04 +1000 Subject: Sharing my ports tree development notes In-Reply-To: <647e7666-c078-4009-9795-defd3503ce70@app.fastmail.com> References: <647e7666-c078-4009-9795-defd3503ce70@app.fastmail.com> Message-ID: <966b2ce7-534b-4d34-a986-6d4ddfb5eade@macports.org> Everything used to be in one repo, so the intent would have been for the doc subdir to be the same license as base and ports. BSD licensing isn't the best fit for documentation but it's adequate. If someone wants a project, contacting all the guide contributors and obtaining permission to relicense under a better documentation license would be one. :) - Josh On 2/9/2026 05:40, Mark E Anderson wrote: > I?m interested in what the license is as well, I have an asciidoc > version that I?m trying to update with how the portgroups work and bring > up to date. Mostly for myself, since I?ve been trying to understand the > internals better but never hurts to share. > > On Tue, Sep 1, 2026, at 10:58 AM, Zero King wrote: >> Hi, >> >> I'm sharing some notes I've written based on my experience contributing >> to the ports tree. Source code for the notes is at >> https://github.com/l2dy/ports-dev-notes > ports-dev-notes> and a generated PDF file is >> available on https://l2dy.github.io/ports-dev-notes/ > l2dy.github.io/ports-dev-notes/> for the time being. >> >> Regarding its license, while I intended to use the same license as >> https://github.com/macports/macports-guide > macports/macports-guide>, I didn't find a license file >> in it. Any suggestions? >> >> Feedback and discussions, either in reply to this email or through >> GitHub issues in my repo, are welcome! >> >> -- >> Best regards, >> Zero King >> > ?Mark > Macports Wiki Page | GitHub > ?| Blog > From macintosh at mathiesen.info Tue Sep 1 21:27:12 2026 From: macintosh at mathiesen.info (Bjarne D Mathiesen) Date: Tue, 1 Sep 2026 23:27:12 +0200 Subject: [boost] v 1.89.0 1.90.0 1.91.0 1.92.0 In-Reply-To: <9bcd431b-a67a-44db-b01f-9f07bb8b1d97@macports.org> References: <52dc1716-d917-4ed3-a09b-1407cfc202bb@mathiesen.info> <9bcd431b-a67a-44db-b01f-9f07bb8b1d97@macports.org> Message-ID: Den 01.09.2026 kl. 23.06 skrev Joshua Root: > On 2/9/2026 05:45, Bjarne D Mathiesen via macports-dev wrote: >> The problem is this : >> ################ >> DEBUG: system: >> cd "/opt/local/var/macports/build/boost189-0610ff64/work/boost_1_89_0" >> && /opt/local/var/macports/build/boost189-0610ff64/work/boost_1_89_0/ >> b2 ... >> runtime-link=shared - j24 architecture=x86 >> ... >> >> notice: could not find main target j24 >> don't know how to make j24 >> ################ > The extra white space is there in the build.args in your Portfiles. Line > 224 in boost189 and the equivalent line in the others. > > - Josh aaaaaarrrrrrggggghh .... -- Bjarne D Mathiesen Kors?r ; Danmark ; Europa ----------------------------------------------------------------------- denne besked er skrevet i et totalt M$-frit milj? MacPro 2010 5.1 ; OpenCore + macOS 15.7.8 Sequoia 2 x 3,46 GHz 6-Core Intel Xeon ; 192 GB 1333 MHz DDR3 ECC RDIMM AMD Radeon Vega VII 16 GB From macintosh at mathiesen.info Mon Sep 7 12:17:54 2026 From: macintosh at mathiesen.info (Bjarne D Mathiesen) Date: Mon, 7 Sep 2026 14:17:54 +0200 Subject: [dovecot 2.4.5] debugging Message-ID: please take a look at https://github.com/macports/macports-ports/pull/34536 I need some pointers regarding finding the root cause -- Bjarne D Mathiesen Kors?r ; Danmark ; Europa ----------------------------------------------------------------------- denne besked er skrevet i et totalt M$-frit milj? MacPro 2010 5.1 ; OpenCore + macOS 15.7.8 Sequoia 2 x 3,46 GHz 6-Core Intel Xeon ; 192 GB 1333 MHz DDR3 ECC RDIMM AMD Radeon Vega VII 16 GB From macintosh at mathiesen.info Mon Sep 7 12:19:35 2026 From: macintosh at mathiesen.info (Bjarne D Mathiesen) Date: Mon, 7 Sep 2026 14:19:35 +0200 Subject: [boost192] debugging Message-ID: <16af14f8-4012-4491-8ce3-203325b8027e@mathiesen.info> please take a look at https://github.com/macports/macports-ports/pull/31402 I need some pointers regarding finding the root cause -- Bjarne D Mathiesen Kors?r ; Danmark ; Europa ----------------------------------------------------------------------- denne besked er skrevet i et totalt M$-frit milj? MacPro 2010 5.1 ; OpenCore + macOS 15.7.8 Sequoia 2 x 3,46 GHz 6-Core Intel Xeon ; 192 GB 1333 MHz DDR3 ECC RDIMM AMD Radeon Vega VII 16 GB From dev at sethm.id.au Fri Sep 11 00:39:25 2026 From: dev at sethm.id.au (Seth McDonald) Date: Fri, 11 Sep 2026 10:39:25 +1000 Subject: Request to test rhino build on macOS 14-15 Message-ID: <7B32ACC4-8FA6-4658-84C1-EABC557EC999@sethm.id.au> Hello all, I recently opened a pull request to update the rhino port (see [0]). Currently GitHub Actions reports the build succeeding for macOS 26 but failing for macOS 14 and 15. However, I suspect these failures may be isolated to the CI environment. So if anyone has access to macOS 14 or 15, would you be able to try building the updated port and letting me know whether it succeeds? And if it fails, could you attach the log? The port uses Gradle to build. The logs of both failed builds attribute the failure to the Gradle daemon being killed early. The relevant log text being: | The message received from the daemon indicates that the daemon has disappeared. | [...] | FAILURE: Build failed with an exception. | | * What went wrong: | Gradle build daemon disappeared unexpectedly (it may have been killed or may have crashed) I suspect this to be isolated to the CI environment because after Googling this error, I found reports of it specifically occurring in CI tests (see [1-2]). The reports typically conclude the daemon was using too much memory and thus it either killed itself due to reaching its memory limit or it was killed by the system due to OOM. I attempted to fix the build failures by adjusting the max memory between 128 MB and 3 GB (hence all the force pushes), but to no avail. Take care, Seth McDonald. [0] [1] [2] -- E9D1 26A5 F0D4 9DF7 792B C2E2 B4BF 4530 D39B 2D51 From macintosh at mathiesen.info Fri Sep 11 05:53:09 2026 From: macintosh at mathiesen.info (Bjarne D Mathiesen) Date: Fri, 11 Sep 2026 07:53:09 +0200 Subject: Request to test rhino build on macOS 14-15 In-Reply-To: <7B32ACC4-8FA6-4658-84C1-EABC557EC999@sethm.id.au> References: <7B32ACC4-8FA6-4658-84C1-EABC557EC999@sethm.id.au> Message-ID: I'll give it a try & report -- Bjarne D Mathiesen Kors?r ; Danmark ; Europa ----------------------------------------------------------------------- denne besked er skrevet i et totalt M$-frit milj? MacPro 2010 5.1 ; OpenCore + macOS 15.7.8 Sequoia 2 x 3,46 GHz 6-Core Intel Xeon ; 192 GB 1333 MHz DDR3 ECC RDIMM AMD Radeon Vega VII 16 GB From macintosh at mathiesen.info Fri Sep 11 08:05:48 2026 From: macintosh at mathiesen.info (Bjarne D Mathiesen) Date: Fri, 11 Sep 2026 10:05:48 +0200 Subject: Request to test rhino build on macOS 14-15 In-Reply-To: References: <7B32ACC4-8FA6-4658-84C1-EABC557EC999@sethm.id.au> Message-ID: <51da4cde-4aa0-4923-954e-cc6cf1a042d7@mathiesen.info> log sent directly -- Bjarne D Mathiesen Kors?r ; Danmark ; Europa ----------------------------------------------------------------------- denne besked er skrevet i et totalt M$-frit milj? MacPro 2010 5.1 ; OpenCore + macOS 15.7.8 Sequoia 2 x 3,46 GHz 6-Core Intel Xeon ; 192 GB 1333 MHz DDR3 ECC RDIMM AMD Radeon Vega VII 16 GB