Re: CI system for PR builds

GitLab 10.6 released with CI/CD for GitHub https://about.gitlab.com/2018/03/22/gitlab-10-6-released/#open-source-projects

Re: CI system for PR builds

On 2 Apr 2018, at 22:53, Mojca Miklavec wrote: > I fail to find any info about available runners for hosted CI infrastructure. Is this what you're looking for? https://about.gitlab.com/features/github/ search for 'self-hosted'.

Re: CI system for PR builds

On 2 Apr 2018, at 23:09, Clemens Lang wrote: > On Mon, Apr 02, 2018 at 08:43:52PM +0200, db wrote: >> GitLab 10.6 released with CI/CD for GitHub > Interesting, but I couldn't find any indication that this would support > builds on macOS. Do you have more info? Nope, bu

Re: CI system for PR builds

On 2 Apr 2018, at 23:09, Clemens Lang wrote: > On Mon, Apr 02, 2018 at 08:43:52PM +0200, db wrote: >> GitLab 10.6 released with CI/CD for GitHub > Interesting, but I couldn't find any indication that this would support > builds on macOS. Do you have more info? I thin

Re: CI system for PR builds

On 3 Apr 2018, at 02:20, Rainer Müller wrote: > Then please explain what this would offer us at all? I'll try. GitLab let's you have external runners (https://docs.gitlab.com/ee/ci/runners/README.html), while TravisCI offers a certain amount of pipeline minutes and only supports the currently s

Re: CI system for PR builds

On 3 Apr 2018, at 13:09, Rainer Müller wrote: > But what exactly do you think would be the benefit from such a complicated > setup (GitHub -> GitLab -> External Runner)? Testing? But hey, I only know MP's infrastructure from what I read in the list, TravisCI does wonders and everything's jolly

Re: CI system for PR builds

On 3 Apr 2018, at 18:04, Mojca Miklavec wrote: > Travis has lots of limitations, but it offers both (a) and (b) for free. Couldn't (b) be the current infrastructure? If I understood correctly previous posts, Travis builds PRs for 10.11-10.13, but then Buildbot builds the binaries that are being

Re: CI system for PR builds

On 7 Apr 2018, at 07:47, Mojca Miklavec wrote: > No, because that would make the infrastructure that distributes binaries to > all our users susceptible to malicious PRs. How is reviewing a PR, then letting it build and, if it succeeds, uploading the binary, if not, reporting the results — diff

Re: CI system for PR builds

Thanks for the explanation. On 7 Apr 2018, at 14:37, Ryan Schmidt wrote: > Only after a PR has been approved and merged to master should a binary be > uploaded anywhere. That's what I meant — but reusing buildbot for testing the PRs. > Our buildbot system was originally set up in 2011 by macOS

Re: CI system for PR builds

On 7 Apr 2018, at 19:44, Clemens Lang wrote: > Remember that Portfiles can execute arbitrary code and root access is > available from Portfiles. We do not want to run arbitrary code in a PR > on the same build machines we use to build packages that we will > distribute to our users. A malicous att

Re: CI system for PR builds

On 8 Apr 2018, at 12:42, Zero King wrote: > On Sun, Apr 08, 2018 at 12:20:34PM +0200, db wrote: >> If you review the code before, that should never be the case and it would >> build just once if it succeeds, right? Or am I missing something how PRs are >> handled? > CI

Re: CI system for PR builds

On 8 Apr 2018, at 13:18, Ryan Schmidt wrote: > On Apr 8, 2018, at 06:12, db wrote: >> That seems to waste at least buildtime and storage. > What do you mean? >> Many updates and revbumps could be almost automatically deployed (after a >> quick review). > Isn't

Re: CI system for PR builds

On 8 Apr 2018, at 13:34, Ryan Schmidt wrote: > No, as I think we've explained several times already, now, Travis CI builds > the proposed change automatically, immediately after it's submitted. The > purpose is to verify that it builds, and the results of that build can inform > our review. Th

Re: CI system for PR builds

On 8 Apr 2018, at 13:50, Ryan Schmidt wrote: > On Apr 8, 2018, at 06:49, db wrote: > On 8 Apr 2018, at 13:34, Ryan Schmidt wrote: >>> No, as I think we've explained several times already, now, Travis CI builds >>> the proposed change automatically, immediat

Re: CI system for PR builds

On 10 Apr 2018, at 20:07, Mojca Miklavec wrote: > On 7 April 2018 at 15:45, db wrote: >> On 7 Apr 2018, at 14:37, Ryan Schmidt wrote: >> Is buildbot running on your basement??? > Yes (not mine). > […] >> Testing and reproducibility, doesn't seem to me as user to be

Re: CI system for PR builds

On 11 Apr 2018, at 15:44, Mojca Miklavec wrote: > On 11 April 2018 at 14:11, db wrote: >> On 10 Apr 2018, at 20:07, Mojca Miklavec wrote: >>> >>>> That streamlined process is what keeps new and updated portfiles in my >>>> local repo… >>> I h

Re: How to submit changes with a GitHub Pull Request (was Re: CI system for PR builds)

On 11 Apr 2018, at 19:24, Perry E. Metzger wrote: > The main steps are: > […] > [By the way, if someone wants to turn this email into a document, I > was pretty careful writing what's above so that would be easy.] Thanks for the write-up. Although I probably rather prefer the PR way, it's gettin

Re: CI system for PR builds

On 8 Apr 2018, at 02:04, Rainer Müller wrote: > So here is the full plan in detail: Side note: Wouldn't it be feasible to adjust portfiles based on CI's feedback? Let's say, you have something build on 10.6-10.13, but for whatever reason it fails on 10.9, so that could be reflected in the portf

Re: How to submit changes with a GitHub Pull Request (was Re: CI system for PR builds)

On 12 Apr 2018, at 14:27, Mojca Miklavec wrote: > What nobody mentioned so far is that one can also simply click "Edit" on the > existing Portfile on GitHub interface and that will also open a pull request. Interesting. It seems that 'Create new file' would have a similar effect. 'Upload files'

Re: CI system for PR builds

On 12 Apr 2018, at 18:52, Ryan Schmidt wrote: > 1. MacPorts does not have a method of declaring that a port does not build on > a version of macOS. Such a feature is being discussed: > https://trac.macports.org/ticket/15712 > In the absence of this feature, we write pre-fetch blocks that manually

Re: How to submit changes with a GitHub Pull Request (was Re: CI system for PR builds)

On 12 Apr 2018, at 19:52, Craig Treleaven wrote: > Is there a playground somewhere to try out such features? For those of us > that are somewhat git- and github-challenged. Not that I'm aware of. Just fork some repo or create a new one and make it your playground.

Re: CI system for PR builds

On 12 Apr 2018, at 20:16, Ryan Schmidt wrote: > You may have noticed that discussion about the issue resumed 4 weeks ago, and > a milestone was assigned. I didn't notice the milestone. Is 2.6 next year's?

Re: macports rot

On 16 Dec 2018, at 13:00, macports-dev-requ...@lists.macports.org wrote: > Message: 1 > Date: Sat, 15 Dec 2018 13:48:21 +0100 > From: Nils Breunese > > I’m also not involved in most upstream projects for the ports I maintain, but > especially when the upstream project has a website project you c

New website design

On 6 Feb 2019, at 13:00, macports-dev-requ...@lists.macports.org wrote: > Message: 2 > Date: Tue, 5 Feb 2019 21:31:22 +0100 > From: Mojca Miklavec > To: MacPorts Development > Subject: New website design > Message-ID: > > Content-Type: text/plain; charset="UTF-8" > Our website design is i

GSoC 2019

On 2 Apr 2019, at 17:11, macports-dev-requ...@lists.macports.org wrote: > GSoC 2019 Could it be possible to relocate all this topics somewhere else?

ungoogled-chromium - Google Chromium, sans integration with Google

Hi I filed a request for this port. I post to the list because I think ungoogled-chromium is noteworthy, regardless if someone tries to make a port for it. https://github.com/Eloston/ungoogled-chromium https://trac.macports.org/ticket/62505

Re: [macports-ports] branch master updated: mosh: 1.3.0, switch to CommonCrypto

On 2 Apr 2017, at 20:42, Clemens Lang wrote: > On Sat, Apr 01, 2017 at 12:48:08PM -0500, Ryan Schmidt wrote: >> The previous version of the port was still building successfully on >> Leopard and later. We may not care about systems so old anymore, but I >> wanted to make sure we were making this d

Re: [macports-ports] branch master updated: mosh: 1.3.0, switch to CommonCrypto

On 3 Apr 2017, at 17:52, Ken Cunningham wrote: > so this sorta gets back to your question about support for older systems, and > is a pretty typical example of same. > sounds like you could write up an openssl variant for it, default to that > variant on suitable systems, and nobody would likely

python port - use pypi or github

When writing a python port, what is best practice for one that's on both PyPI and GitHub, use pypi2port or portgroups github+python?

Re: python port - use pypi or github

On 10 Apr 2017, at 12:59, Clemens Lang wrote: > It's up to you, really. pypi2port might be easier, because it auto-generates > most > of the boilerplate and is better tested, since most python ports use PyPI. Ok, I just want to avoid having too much divergence between current development, relea

highlight #53994 #53719 #53542

These tickets could be closed. It builds fine with clang-3.9 and libc++ on 10.8 and possibly older.

pypi2port and python supported versions

I used pypi2port to get the basic porfile for stem module, emended and submitted it in #54014. I was told to remove py34 and add py36. pypi2port usually writes 27 34 35 in python.versions, although 33 is still supported (now for a couple of months) and 36 is already released. What is the genera

Re: pypi2port and python supported versions

On 22 Apr 2017, at 23:17, Mojca Miklavec wrote: > 27 35 36 for sure. > 34 has not yet seen any active removal, so it's the bordercase, > probably not worth putting into new ports. These are the EoS dates: 3.3 2017-09 3.4 2019-03 I agree on 33, but 34? Why don't we also leave 35 out?

Re: pypi2port and python supported versions

On 24 Apr 2017, at 19:10, Mojca Miklavec wrote: > In short: > - we probably need a bit more effort to bring all modules to 3.6 > - we probably need to review which entries may go from py-graveyard > and potentially change the others to default to 36 rather than 34 > - we certainly need to upgrade

env python - port select or reinplace

What should I use to reference the python binary, put notes to use port select or rewrite the interpreter directive?

Re: env python - port select or reinplace

On 27 Apr 2017, at 19:11, Joshua Root wrote: > If you have a dependency on a pythonXY port you need to make the scripts use > that particular version. I have #! /usr/bin/env python3 in the scripts, no particular version. But even if I had one, and it run on any (recent) version, what should I u

Re: env python - port select or reinplace

On 28 Apr 2017, at 03:19, Ryan Schmidt wrote: > Use the latest stable version of python. Or offer variants so the user can > choose, and default to the latest stable version. Could you please take a look at #54054? The port rewrites some paths before copying the files in place — nothing fancy.

Re: Python 3.4

On 5 May 2017, at 23:35, Mojca Miklavec wrote: > - Should new ports include 34 subports or not? > - Should we actively remove 34 subports? Judging by PEP 398 for v3.3 and PEP 494 for v3.6, v3.4 will live until around 03/2019 (no statement in its PEP 429). > I have nothing against leaving suppor

Re: Python 3.4

On 8 May 2017, at 18:36, m...@macports.org wrote: >> - Should we actively remove 34 subports? > When py35 and py36 have more sub-ports, then yes. I wonder how many of those are maintained, of these how many actually get updated and how many are non-maintained at all.

Re: mDNSResponder building

On 26 Jun 2017, at 18:03, René J.V. Bertin wrote: > Can anyone here tell me how to build the mDNSResponder (so that I can stop it > polluting my log with "mDNSResponder[43]: Unknown DNS packet type 6F74 from > 192.168.1.14 :57621 to 192.168.1.255 :57621 length 44 on 0004 > (ignor

tarball link when using python group

I made a vanilla port for textfsm with pypi2port, but want to use the tarball from github. curl will download https://codeload.github.com/google/textfsm/tar.gz/0.3.2 without the proper suffix. How can I fix it? I tried fetch.type, but it seems for another use case.

install doc with python group

Which is the proper way to install documentation with python group? This boilerplate works for the subport but not if I install the main port. post-destroot { set docdir ${prefix}/share/${subport} xinstall -d ${destroot}${docdir} xinstall -m 644 -W ${worksrcpath} \ CHANGES.txt

registry with multiple portfiles

Is it normal or expected to have multiple portfiles for the same port in the registry?

Re: install doc with python group

On 11 Jul 2017, at 14:44, Joshua Root wrote: > Only the subports like py27-setuptools_hg and py35-setuptools_hg are supposed > to install any files. py-setuptools_hg is a stub port, and as such has no > distfiles and thus no extracted sources. I know why it fails. The stub port does install the

Re: registry with multiple portfiles

On 11 Jul 2017, at 14:50, Joshua Root wrote: > On 2017-7-11 21:09 , db wrote: >> Is it normal or expected to have multiple portfiles for the same port in the >> registry? > When multiple versions of the port are installed, yes. And when having just one? $ tree /opt/local/var

Re: registry with multiple portfiles

On 11 Jul 2017, at 16:16, Joshua Root wrote: > How do the files differ? Is one not referenced in the registry database? port cat shows the right file. I think I might explain this case — but not others. While writing the portfile for py-ipaddr in vim I saved it and stopped the process, then ins

Re: registry with multiple portfiles

On 11 Jul 2017, at 19:21, Joshua Root wrote: >> I think I might explain this case — but not others. While writing the >> portfile for py-ipaddr in vim I saved it and stopped the process, then >> installed/uninstalled the port and made some corrections in the portfile, >> repeat. Does it sound p

Re: tarball link when using python group

On 9 Jul 2017, at 23:29, Joshua Root wrote: > Solution 2: Use the github portgroup, which takes care of the idiosyncrasies > of fetching from github. It's fine to use at the same time as the python > portgroup, see e.g. py-alabaster. Thanks for the reference. I was missing setting name to py-mo

Re: install doc with python group

On 11 Jul 2017, at 15:45, Rainer Müller wrote: > Wrap your code in a section that is only executed when installing the subport: > if {${subport} ne ${name}} { >post-destroot { D'oh! Of course that did it. Thanks for point that out.

Re: registry with multiple portfiles

On 11 Jul 2017, at 18:40, db wrote: > While writing the portfile for py-ipaddr in vim I saved it and stopped the > process, then installed/uninstalled the port and made some corrections in the > portfile, repeat. I tried this with another port while monitoring its registry fold

Re: registry with multiple portfiles

/registry/portfiles/ ./find_unreferenced_portfiles.tcl | while read LINE ; do sudo rm -r $MPREGPF$LINE ; done ls $MPREGPF | \ while read LINE ; \ do [[ $(ls $MPREGPF$LINE | wc -l | xargs) == 0 ]] \ && sudo rm -r $MPREGPF$LINE ; done Is it safe to delete the rest? I haven't manually checked the reg db yet.

Re: registry with multiple portfiles

On 13 Jul 2017, at 03:37, Joshua Root wrote: > I don't know. Theoretically if the script doesn't print something then a row > in the ports table matched it. Here's an extract of my filter, reg portfiles dir + db row(s). It doesn't make sense as port dir it's di

Re: registry with multiple portfiles

On 13 Jul 2017, at 18:08, Joshua Root wrote: > Fixed in master: > I understand that there's no additional name field for subports in the registry and that they have the same hash. But I don't get how the

Re: registry with multiple portfiles

On 13 Jul 2017, at 21:50, db wrote: > On 13 Jul 2017, at 18:08, Joshua Root wrote: >> Fixed in master: >> <https://github.com/macports/macports-base/commit/6c8c70b370753b9015400b27afdcd303b7a0cb63> > I understand that there's no additional name field for subport

Re: registry with multiple portfiles

On 17 Jul 2017, at 10:38, Joshua Root wrote: > Don't delete anything from the database, its contents are correct (and have > been the whole time). You can delete the portfiles that are in a directory > named after a port (or version of a port) that is not installed. Those should > have been del

Re: [GSoC] migration

On 20 Jul 2017, at 06:04, Bradley Giesbrecht wrote: > Snapshot and restore should be able to be executed without migrate. I have a question and I haven't been following in detail this GSoC, so bear with me. Could snapshot and restore be used to rollback to a previous state in case a sync and u

Re: [GSoC] migration

On 21 Jul 2017, at 13:02, Umesh Singla wrote: > Unless we have a snapshot of the previous state, that is, before it got > hampered. > But then again, we reinstall all the ports presently. At this time, it could > be hard for me to detect what went wrong while sync or upgrade. If I understood yo

Re: [GSoC] migration

On 22 Jul 2017, at 03:01, Umesh Singla wrote: > I don't know, in the above example, what do you mean when you say "..you > realize that from its deps..", like, realize how? I am just asking this if > there's some other way to get info on the latest modifications that I might > not be aware of.

Re: [GSoC] migration

On 27 Jul 2017, at 00:35, Umesh Singla wrote: > Hi Josh >> We don't have version dependencies so no, this sort of check is not possible >> a priori. Breakage will be caught by rev-upgrade after the fact in many >> cases (and if it is set to rebuild automatically, it may well revert you >> back

Re: [GSoC] migration

On 27 Jul 2017, at 21:21, Bradley Giesbrecht wrote: > Migration will uninstall all installed ports and then install all the ports > from the snapshot Yes, but I'm talking about using snapshot and restore at different times with ports being updated between them, likely resulting in different out

recursively copy docs

I want to recursively copy a port's documentation which has nested dirs. Is there any caveat for using exec cp -R ${worksrcpath}/sphinx-doc ${destroot}${docdir} instead of the more canonical form in the wiki? https://trac.macports.org/wiki/PortfileRecipes#doc

Re: recursively copy docs

On 29 Jul 2017, at 20:39, Ryan Schmidt wrote: > copy ${worksrcpath}/sphinx-doc ${destroot}${docdir} That macro works, but file copy errored each time I tried it before.

reinplace first occurrence

I checked portfile's man page and peaked at the source, but couldn't figure out yet, why these forms don't work. https://stackoverflow.com/a/11458836/2167331 In the meantime, I worked around it, but I'd still like to know how to do it.

Re: reinplace first occurrence

On 3 Aug 2017, at 18:46, Ryan Schmidt wrote: > You could run a two-command sed manually (with system, not reinplace). Or > maybe you can use awk, or an ed script. This works. Caveat emptor: || (empy RE) doesn't work and range seems to start on first line, not on its subsequent. reinplace -E 1,

get user's home

How can I get the user's home, the flesh and blood bot's? set home$::env(HOME) righteously returns /opt/local/var/macports/home/

Re: get user's home

On 4 Aug 2017, at 16:05, Joshua Root wrote: > Should be in $user_home (if it could be determined, which is not guaranteed). > But why do you need it? The user installing the port is not necessarily the > one that will use it. That worked, thanks. I used it for an env var's default value in ipfs

Re: get user's home

On 4 Aug 2017, at 22:58, Joshua Root wrote: > Suppose a port is built on the buildbot. The user is 'buildbot' and so the > value of $user_home will be something like '/Users/buildbot'. During testing I didn't thoroughly checked that env var there. I suppose, it could be deleted or set to nil an

Re: get user's home

On 5 Aug 2017, at 10:00, Joshua Root wrote: > I don't think there is a good way to do it if the user's home directory must > be encoded into the plist. But I suspect that is not the case, and Rainer has > mentioned a good alternative in the ticket. IPFS_PATH sets the path to ipfs' repo for file

Re: get user's home

On 7 Aug 2017, at 20:03, Bradley Giesbrecht wrote: >> On Aug 5, 2017, at 3:55 AM, db wrote: >>> docker-machine's port, although I don't quite grasp what depends_run-append >>> exactly does there. > Depends_run-appends adds to the depends_run variable. T

Re: get user's home

On 7 Aug 2017, at 23:03, Joshua Root wrote: > Dependency specifications are documented here: > I already read that too, but don't see why in case of docker-machine a path dependency makes more sense th

Re: get user's home

On 8 Aug 2017, at 02:12, Ryan Schmidt wrote: > path:-style dependencies are used when more than one port could satisfy the > dependency, in this case the development and stable versions of > bash-completion. Thanks, that cleared it up. Re ipfs, I left the startup item for the user to install

Re: Why does macports require a migration across major version bumps?

On 30 Aug 2017, at 20:29, Clemens Lang wrote: > Use 'sudo port -f selfupdate' instead, which should also recompile your > version of MacPorts. > In theory, reinstalling MacPorts and using 'sudo port -t upgrade > outdated' after an OS update should thus have the same effect as the > migration inst

Re: Switching cxx_stdlib for C++11 on legacy macOS

On 3 Sep 2017, at 13:47, Rainer Müller wrote: > The other option to switch would be to just change the default of > cxx_stdlib with the next 2.x release, wipe all existing binary packages > and ask all existing users on legacy systems to uninstall, then > reinstall all their ports. Less work for d

Re: Switching cxx_stdlib for C++11 on legacy macOS

On 4 Sep 2017, at 20:16, Ken Cunningham wrote: > switch to libc++: > Benefits: > 1. most builds are identical to 10.9+, most portfiles need no changes. > Drawbacks: > 1. needs changes in macports infrastructure to support new buildbots and > pre-built binaries > 2. poorly-written software or Port

Re: Restoring from Time Machine backup relocates home directories

On 15 Sep 2017, at 04:32, Ryan Schmidt wrote: > I recently reinstalled macOS and, during the setup process, restored from a > Time Machine backup. One would think that after restoration, the machine > would be in a state sufficiently similar to the one it was in at backup time, > but for our pu

Re: Restoring from Time Machine backup relocates home directories

On 15 Sep 2017, at 18:06, Ryan Schmidt wrote: > Not always, because many ports use local variables when setting the home > directory, so you then have to find the place in the Portfile where that > local variable is set. And also, many ports split the add_users invocation > over multiple lines,

Re: Restoring from Time Machine backup relocates home directories

On 15 Sep 2017, at 23:47, Ryan Schmidt wrote: > MacPorts and ports can add users, but MacPorts doesn't remove them. I'm not > sure how we could change that. Suppose you install a port that has a user > that needs to own some data, such as postresql96-server. You then uninstall > the port, which

Re: Restoring from Time Machine backup relocates home directories

On 16 Sep 2017, at 16:03, Rainer Müller wrote: > As you just run it on your system, you know how long that will take. > Clearly not an option to do this before/after every install/uninstall. Only those that create users. > And what would you gain from that? A free uid. They are not a scarce > re

Re: Restoring from Time Machine backup relocates home directories

On 17 Sep 2017, at 00:15, Ryan Schmidt wrote: > File ownership can be preserved on disk images, but that ownership is > preserved by uid and gid, not by user name and group name. I suspected this but didn't know how it actually works on OS X. >> MP could also checked if the UIDs/GIDs are alread

Re: virtual machine for 10.5 intel ?

On 3 Oct 2017, at 05:01, Ken Cunningham wrote: > if anyone would care to enlighten me on the best approach to making a VM for > 10.5 intel, I would appreciate it. This is the only system I have that I > can’t presently see a way to virtualize. At that time I wanted to do the same and found this

Re: Releasing 2.4.2

On 6 Oct 2017, at 18:54, Ryan Schmidt wrote: > Thanks, that was exactly the fix I was just going to suggest. I'll deploy > that, then 2.4.2 should be available within the hour. Sorry to hijack the thread. I just installed 2.4.2 from source and realised that rleaves didn't make it to it yet. Any

Re: State of the GnuPG ports

On 10 Oct 2017, at 07:18, Leonardo Brondani Schenkel wrote: > On 2017-10-10 03:10, Helmut K. C. Tessarek wrote: >> On 2017-10-09 08:11, Rainer Müller wrote: >>> gnupg -> 2.2 >>> gnupg-devel -> 2.3 >>> gnupg1 -> 1.4 >> +1 > +1 +1

Re: Force build on wrong builder

On 12 Oct 2017, at 14:37, Ryan Schmidt wrote: > The High Sierra builder has been busy building ports as commits come in. It > builds and deploys binaries not only for the port being modified, but also > for all of its dependencies recursively. Does a commit to a dependency trigger a rebuild of

Re: Force build on wrong builder

On 12 Oct 2017, at 16:13, Ryan Schmidt wrote: > On Oct 12, 2017, at 09:11, db wrote: >> >> Does a commit to a dependency trigger a rebuild of its rdependents? > Unfortunately no. We've been discussing that on the -infra list but I don't > think we found a solutio

Re: Force build on wrong builder

On 12 Oct 2017, at 22:23, Joshua Root wrote: > All old port versions are uninstalled after each build, so the deps will > never be outdated during a build. Also a build would have been triggered in > response to any earlier commit that updated one of the deps. Let's say, port A depends recursiv

base master and editing commands

Basic editing commands don't work in shell mode. macports-base $ port MacPorts 2.4.99 Entering shell mode... ("help" for help, "quit" to quit) [Code/macports-base] > syncc^H^H^A^E^K Fortunately ^U works. I haven't found an issue in the tracker.

Re: base master and editing commands

On 4 Dec 2017, at 00:37, Joshua Root wrote: > Looks like you have a build without readline support? Hmm…, I just cloned base master and built it. Only option I added is --with-curlprefix=/opt/local to the configure phase.

Re: base master and editing commands

On 4 Dec 2017, at 01:00, Joshua Root wrote: > So you want --enable-readline as well. Thanks. In the guide 2.2.2. doesn't use this option, 2.2.3. does.

Re: base master and editing commands

On 4 Dec 2017, at 01:00, Joshua Root wrote: > So you want --enable-readline as well. I just tried it and some of the commands still don't work: ^A, ^E, arrows — ^U, ^W work though.

Re: base master and editing commands

On 4 Dec 2017, at 01:00, Joshua Root wrote: > So you want --enable-readline as well. I tried cleaning and rebuilding, and now I get the following error. readline.c:128:23: error: use of undeclared identifier 'username_completion_function'; did you mean 'rl_username_completion_function'?

Re: base master and editing commands

On 4 Dec 2017, at 13:20, Rainer Müller wrote: > You are trying to build MacPorts against readline installed by MacPorts. Why can I build it against its own curl but not its own readline? Eventually, I temporarily deactivated readline and it built fine.

Re: Too many compilers needed for C++11

On 6 Dec 2017, at 00:46, Ken Cunningham wrote: > clang/llvm 5.0 is the current cxx11 1.1 compiler, so that is what is being > called up. Shouldn't such portgroup dependencies be listed by `port rdeps`? I don't see it, for example, in highlight.

Re: Too many compilers needed for C++11

On 7 Dec 2017, at 12:59, Mojca Miklavec wrote: > Nowadays many ports claim >Warning: All compilers are either blacklisted or unavailable; defaulting > to first fallback option > but do install nevertheless at the end. I would say that this is a bug in > either the portfiles or portgroups tha

Re: Too many compilers needed for C++11

On 7 Dec 2017, at 07:29, Ryan Schmidt wrote: > They are, if you're running on an OS version where those dependencies are > added. Here's what I see: $ sw_vers ProductName:Mac OS X ProductVersion: 10.8.5 BuildVersion: 12F2560 $ port rdeps highlight The following ports are dependencies of h

Re: Too many compilers needed for C++11

On 7 Dec 2017, at 14:51, Joshua Root wrote: > Looks correct to me; the system clang will be used for C++11 on 10.8 with > libc++ (assuming most recent possible Xcode). Ok. Out of curiosity, I tried on a 10.8-VM that's not set to build from source, with CLT, no Xcode — to install highlight and i

ansible replacement by py-ansible

I tried upgrading ansible from 2.3.2.0.1_1 to 2.3.2.0.1_2 and it failed. The log showed a conflict with ansible (sic) so I cat'ed the portfile and noticed the ui_error in the pre-configure phase stating its replacement by py-ansible and some instructions. So I did install py-ansible and it faile

Re: ansible replacement by py-ansible

On 14 Dec 2017, at 16:53, Joshua Root wrote: > It needs the attached change. Should I reopen #55229 and attach the diff?

Re: ansible replacement by py-ansible

On 15 Dec 2017, at 05:25, Andrew Fernandes wrote: > patched and committed in f5a44cbcb1 Weren't there also man pages?

Re: ansible replacement by py-ansible

On 15 Dec 2017, at 15:41, Andrew Fernandes wrote: > Not in the pypi dist! That's great. I guess, I'll have to make my own or look somewhere else.

reclaim and rleaves

Shouldn't reclaim use rleaves instead of unrequested? As of now I have to rerun reclaim in order to remove unnecessary ports.

  1   2   >