Opened 4 years ago
Last modified 4 years ago
#61557 assigned defect
neomutt @20200925_0 w3m help / header cache problem / no authenticators
Reported by: | arappe | Owned by: | lbschenkel (Leonardo Brondani Schenkel) |
---|---|---|---|
Priority: | Normal | Milestone: | |
Component: | ports | Version: | |
Keywords: | Cc: | l2dy (Zero King), tristancollins (Tristan Collins) | |
Port: | neomutt |
Description
First, thank you for maintaining NeoMutt!
- w3m: w3m does not compile on Big Sur. But I think that neomutt does not rely on w3m. One can use lynx. I edited the PortFile and can compile 20200925 and 20200814 on Big Sur after deleting w3m as a build dependency.
- hcache: I think that the 20200925 update breaks the header cache. Executed on either Catalina and Big Sur, and compiled on either Catalina or Big Sur, 20200925 (specifically neomutt @20200925_0+doc+gdbm+gnutls+gpgme+idn+mutt+sasl) upon startup scans the header cache and then proceeds to reject it and request headers starting from 0. Killing neomutt, activating neomutt @20200814_0+doc+gdbm+gnutls+gpgme+idn+mutt+sasl and then upon startup neomutt scans the same stored header cache file and accepts it, only acquiring the few new headers since previous run.
- No authenticators found. For neomutt compiled on Big Sur for both 20200925 and 20200814 I cannot send outgoing mail. I am told "No authenticators found." The same .muttrc file and the same email account can send outgoing mail in Catalina. (Multiple email accounts: yahoo, gmail, O365) And the "Darwin 19" neomutt compiled on Catalina (which still works on Big Sur) can send outgoing mail in Big Sur (either 20200925 or 20200814). The problem is with Big Sur compilation leading to loss of outgoing mail service. Any suggestions?
Thanks!
Change History (11)
comment:1 Changed 4 years ago by jmroot (Joshua Root)
Cc: | l2dy added |
---|---|
Owner: | set to lbschenkel |
Port: | neomutt added |
Status: | new → assigned |
comment:2 Changed 4 years ago by lbschenkel (Leonardo Brondani Schenkel)
comment:3 Changed 4 years ago by arappe
Thanks for replying. About reporting 2 and 3 to upstream: (what is "upstream" exactly?) Do you mean to the maintainers of mutt (outside of MacPorts)? Can you tell me where/how I do that? I am willing to make the report.
About 1: within this trac system several have reported that w3m does not compile in Big Sur. The reply seems to be that w3m is unmaintained. So even if w3m is better than lynx (interesting by the way) requiring w3m at build time means "no neomutt for me". I thought that reporting the work-around was valuable. And we can continue to hope that someone fixes w3m.
comment:4 Changed 4 years ago by arappe
Updates
- w3m build now works. I see online that clang-9.0 is being used to facilitate compilation. So this issue can be set aside.
- header cache works. Upon switch to Big Sur-compiled neomutt, the header cache is discarded...but only once. After getting the header cache with the Big Sur build, it is then respected by neomutt the next time that neomutt is run.
- I think that the "no authenticators found" is, sadly, an issue for a MacPorts maintainer.
sudo port install neomutt +doc+gdbm+gnutls+gpgme+idn+mutt+sasl still gives an install that cannot send outgoing mail.
However, a standalone compilation of mutt-1.14.6.tar.gz with
configure -with-ssl=/opt/local --with-gdbm=/opt/local --prefix=/Users/xxxxx/DEV/code/software/macosx/muttdir/standalone_compile/mutt_install --enable-hcache --enable-gpgme --with-sasl --enable-imap
sudo gmake install
gives me a version of mutt on the same exact computer and with the same exact muttrc file that can send outgoing mail.
I tried many versions of the MacPorts variants:
1) neomutt @20200925_0
2) neomutt @20200925_0+doc+gdbm+gnutls+gpgme+gss+idn+mutt+sasl
3) neomutt @20200925_0+doc+gdbm+gnutls+gpgme+idn+mutt+sasl
4) neomutt @20200925_0+doc+gdbm+gnutls+gpgme+mutt+sasl
5) neomutt @20200925_0+doc+gdbm+gpgme+idn+mutt+sasl
6) neomutt @20200925_0+doc+gpgme+idn+mutt
7) neomutt @20200925_0+doc+gpgme+idn+mutt+sasl
8) neomutt @20200925_0+gnutls+gpgme+mutt+sasl
9) neomutt @20200925_0+gpgme+mutt+sasl
10) neomutt @20200925_0+mutt
11) neomutt @20200925_0+mutt+sasl
If there's some other combination that I should try, I'd be glad to know about it. Other suggestions?
comment:5 Changed 4 years ago by lbschenkel (Leonardo Brondani Schenkel)
Thanks for the update. I'll see when I can investigate number 3, but as I said given my current personal circumstances I don't have much free time to dedicate to this project. Any investigation and/or information you can gather up in the meantime will be greatly appreciated.
comment:6 Changed 4 years ago by arappe
Thanks. Not rushing you; I totally respect the time you have already put into maintaining neomutt for us!
I now have what I think is a totally working build (outside MacPorts) for neomutt as well as for mutt. So I think it must be the way the neomutt MacPorts variants are translating into the configure flags for the original distribution.
prompt% tar xvfz neomutt-20201120.tar.gz
prompt% cd neomutt-20201120
prompt% gmake clean
prompt% ./configure --gpgme --gnutls --sasl --ssl --gdbm --idn2 --disable-idn
prompt% sudo gmake
(gmake install didn't work in terms of installing in my preferred directory. Tried --prefix and DESTDIR. No luck.) (just run program from build directory)
I report this for other neomutt fans out there, that there is a workaround that's pretty accessible.
comment:7 Changed 4 years ago by tristancollins (Tristan Collins)
Cc: | tristancollins added |
---|
comment:8 Changed 4 years ago by tristancollins (Tristan Collins)
I'm having the same issue. I recently upgraded from (A) Catalina -> to (B) Big Sur but kept MacPorts as it was from (A). No migration or updates. The neomutt binary from (A) worked fine on Big Sur.
I updated MacPorts today and got the error described in (3) above. Details below.
I copied the 'working' neomutt binary from the (A) Catalina backup to the (B) Big Sur system, and it fails just as in (3).
The differences in the outputs from neomutt -v seem to be the version of OpenSSL used (g->h), the Darwin version, the CFLAGS used and the notmuch library.
So perhaps the issue is with a change in OpenSSL?
The output of neomutt -v from a working (A) Catalina 20200814 version:
System: Darwin 19.6.0 (x86_64) ncurses: ncurses 6.2.20200212 (compiled with 6.2.20200212) libiconv: 1.16 libidn2: 2.3.0 (compiled with 2.3.0) GPGME: 1.14.0-unknown OpenSSL: OpenSSL 1.1.1g 21 Apr 2020 libnotmuch: 5.2.0 storage: gdbm compression: zlib Configure options: --prefix=/opt/local --enable-doc --disable-idn --with-ncurses=/opt/local --with-nls=/opt/local --with-ssl=/opt/local --with-zlib=/opt/local --with-gdbm=/opt/local --with-gpgme=/opt/local --with-idn2=/opt/local --with-notmuch=/opt/local --with-sasl=/opt/local Compilation CFLAGS: -pipe -Os -isysroot/Library/Developer/CommandLineTools/SDKs/MacOSX10.15.sdk -arch x86_64 -std=c99 -D_ALL_SOURCE=1 -D_GNU_SOURCE=1 -D__EXTENSIONS__ -I/opt/local/include -DNCURSES_WIDECHAR Default options: +attach_headers_color +compose_to_sender +compress +cond_date +debug +encrypt_to_self +forgotten_attachments +forwref +ifdef +imap +index_color +initials +limit_current_thread +multiple_fcc +nested_if +new_mail +nntp +pop +progress +quasi_delete +regcomp +reply_with_xorig +sensible_browser +sidebar +skip_quoted +smtp +status_color +timeout +tls_sni +trash Compile options: -autocrypt +bkgdset +color +curs_set +fcntl -flock -fmemopen +futimens +getaddrinfo -gnutls +gpgme -gss +hcache -homespool +idn -inotify -locales_hack -lua +meta -mixmaster +nls +notmuch +openssl +pgp +regex +sasl +smime -sqlite +start_color +sun_attachment +typeahead MAILPATH="/var/mail" PKGDATADIR="/opt/local/share/neomutt" SENDMAIL="/usr/sbin/sendmail" SYSCONFDIR="/opt/local/etc"
From a non-working BigSur 20200925 MacPorts version:
System: Darwin 20.1.0 (x86_64) ncurses: ncurses 6.2.20200212 (compiled with 6.2.20200212) libiconv: 1.16 libidn2: 2.3.0 (compiled with 2.3.0) GPGME: 1.14.0-unknown OpenSSL: OpenSSL 1.1.1h 22 Sep 2020 libnotmuch: 5.3.0 storage: gdbm compression: zlib Configure options: --prefix=/opt/local --enable-doc --disable-idn --with-ncurses=/opt/local --with-nls=/opt/local --with-ssl=/opt/local --with-zlib=/opt/local --with-gdbm=/opt/local --with-gpgme=/opt/local --with-idn2=/opt/local --with-notmuch=/opt/local --with-sasl=/opt/local Compilation CFLAGS: -pipe -Os -isysroot/Library/Developer/CommandLineTools/SDKs/MacOSX11.0.sdk -arch x86_64 -std=c99 -D_ALL_SOURCE=1 -D_GNU_SOURCE=1 -D__EXTENSIONS__ -I/opt/local/include -DNCURSES_WIDECHAR Default options: +attach_headers_color +compose_to_sender +compress +cond_date +debug +encrypt_to_self +forgotten_attachments +forwref +ifdef +imap +index_color +initials +limit_current_thread +multiple_fcc +nested_if +new_mail +nntp +pop +progress +quasi_delete +regcomp +reply_with_xorig +sensible_browser +sidebar +skip_quoted +smtp +status_color +timeout +tls_sni +trash Compile options: -autocrypt +bkgdset +color +curs_set +fcntl -flock -fmemopen +futimens +getaddrinfo -gnutls +gpgme -gss +hcache -homespool +idn -inotify -locales_hack -lua +meta -mixmaster +nls +notmuch +openssl +pgp +regex +sasl +smime -sqlite +start_color +sun_attachment +typeahead MAILPATH="/var/mail" PKGDATADIR="/opt/local/share/neomutt" SENDMAIL="/usr/sbin/sendmail" SYSCONFDIR="/opt/local/etc"
I can't get the below sequence to result in a 'working' version on my system for issue (3) above:
prompt% gmake clean prompt% ./configure --gpgme --gnutls --sasl --ssl --gdbm --idn2 --disable-idn prompt% sudo gmake
comment:9 follow-up: 10 Changed 4 years ago by arappe
As of today, I find that the MacPorts neomutt with BigSur has the same "no authenticators" problem as above. No change. Clean install of MacPorts. sudo port install neomutt (with flags as above) --> "no authenticators".
Therefore, the only way I can be helpful maybe is in figuring out why your standalone neomutt compile (from .tgz) didn't work.
You wrote "I can't get the below sequence to result in a 'working' version on my system for issue (3) above:
(do you mean it didn't compile? didn't run? still complains about no authenticators?) Post some more info on that sub-issue and maybe I can get you back on the road. And perhaps knowing how the standalone version can compile will make it easier for someone MacPorts-savvier than me to adapt the MacPorts version later on....
comment:10 Changed 4 years ago by tristancollins (Tristan Collins)
Replying to arappe:
You wrote "I can't get the below sequence to result in a 'working' version on my system for issue (3) above:
(do you mean it didn't compile? didn't run? still complains about no authenticators?) Post some more info on that sub-issue and maybe I can get you back on the road. And perhaps knowing how the standalone version can compile will make it easier for someone MacPorts-savvier than me to adapt the MacPorts version later on....
Still complains about the no authenticators. I ended up up switching to Homebrew and using the version there - works fine.
comment:11 Changed 4 years ago by arappe
Just a quick post to say that I upgraded to OSX 11.2.3 and Xcode 12.4 (I get the complaint that port cannot tell what version of the command line tools I have, despite my following all the hotlist suggestions)
With this OS X config I find that neomutt installs properly and can cache headers and authenticate. AFAICS neomutt in MacPorts now works as advertised. Not sure what changed; just wanted people to know.
% sudo port install neomutt +doc+gdbm+gnutls+gpgme+idn+mutt+sasl
% port installed|grep neomutt
neomutt @20200925_0+doc+gdbm+gnutls+gpgme+idn+mutt+sasl (active)
Just making a report that the problems I saw before seem to have worked out.
Would you be willing to report 2 and 3 to upstream? I could do that, but I'm not sure if there's any advantage of me being an intermediary.
Regarding 1, I no longer remember with confidence but I have the impression that I tried all options, and
w3m
was the only one that rendered the doc faithfully. The others produced docs with visual defects. (But then again, I don't remember the details.) This doesn't mean this can't be re-evaluated. I must warn, though, that due to personal circumstances I'm quite short on free time at the moment.