Hello, all

Could I ask for some help with darcs here?

I feel to being unable to rename a depended-upon patch in my darcs repo. Is it possible at all?

I think the rebase command is the way

I use `darcs amend --edit-long-comment -h <some_hash> -m <some-message>`

And got "Skipping depended-upon patch: <the_hash>." … And then "Cancelling amend since no patch was selected."

How should I use rebase for it?

Sorry, I don’t remember but is there a manual for it ?

(I'm a very new darcs user and the actual task I'm trying to do is converting my old git repos with commit messages containing cyrillic characters. I was successful with direct `git-fast-export | darcs convert import` route as the resulting patch messages were somewhat incorrectly encoded. Then I decided to write a shell script, calling git-filter-repo with Python callback, base-64 encoding all the commit messages in-place before

export, and a small program in Elixir, which should read and parse `darcs log --xml-output` and then issue `darcs amend` command for each comment with base64-decoded patch name)

This is the last step in my path, and I feel being unable to solve this without the help (or some advice) from some darcs developers…

Seems there should be some simplier way to do this. But I still didn't find any API to darcs to operate on a repo as a whole (somewhat like `git-filter-branch`).

korobkov: I see ! would it be an option to fix up all the names in git first, where the tools are much more powerful ?

Yes :) You are right. And it would be much more simple for me if it would be possible for the names to be correctly imported by darcs!

(If I create cyrillic messages from scratch they are listed in `darcs log` and exported all OK!)

Seems the issue is only with `darcs convert import`…

hmm, I would research in the manual for anything encoding related then - I remember some environment variable..

I've read that it's not all that simple with UTF-8 in Haskell, but as it already works pretty well with newly created patch names, the internal representation of patches should support it already.

DONT_ENCODE_8_BIT?

(or alike)

Does it also affect UTF-8 somehow?

korobkov: I don't know.. also I assume you have a correct LANG set

% env | grep LANG

LANG=ru_RU.UTF-8

LANGUAGE=

the darcs manual is no longer online, that's unfortunate

darcs help manpage or markdown produces it, and I see a lot of env vars there, is that what you're reading ?

I wonder if DARCS_DONT_ESCAPE* help

Thank you, I'll try…

I don't see any DONT_ENCODE there

maybe those only affect terminal display, not sure

Heffalump might know more.. good luck!

Thank you once again. Hope he could help with it.

Should I ask him here or elsewhere (e.g. via email)?

How do people get answers if they are from different timezones? (I'm somewhat new to IRC…)

korobkov: he'll see it here in due course

That,s OK. :) I'm concerned about viewing channel logs at tomorrow morning, should I have any reboot at night… (as I've heard the logs arebt provided online due to GDPR…)

ah, not this channel has a decent log - see the /topic

s/not/no/

currently still unable to connect to freenode.logbot.info/darcs …

korobkov: that's unfortunate, it works fine here

another option is to connect via the Riot matrix client, which is good at tracking history and notifications

or just ask again next time you're online, he'll probably be around

and/or, search/ask the darcs-users mail list

It may be very likely due to all kinds of internet restrictions in russia in general… :(

:(

I would ask next time in case of disconnect

sounds good

Just tried `git fast-export master | ( setenv DARCS_DONT_ESCAPE_8BIT 1 ; darcs convert import <path> )`. No luck

Just tried `git fast-export master | ( setenv DARCS_DONT_ESCAPE_ANYTHING 1 ; darcs convert import <path> )`. No luck either

It all looks like * РазвÑ<U+0091>Ñ<U+0080>Ñ<U+0082>Ñ<U+008B>вание пеÑ<U+0080>еведено на Ñ<U+0081>Ñ<U+0085>емÑ<U+0083> :remote_cache по ipv6 (без пÑ<U+0080>окÑ<U+0081>и)

Double encoded?

korobkov: I see the same

I see related issues at bugs.darcs.net/issue?%40columns=id%…=convert&submit=Search+issue+titles

probably it needs fixing, posting on some of those issues might get a fix from bf ("I have a patch for this and a few other convert related issues. Stay tuned.")

So, some hope for me? :)

(Either way I'm not about giving up with using Darcs! :) I'll just will use it for all my new repos. (Too sad I didn't know much about Darcs 10 years ago… I for sure would chose it instead of git))

how long have you been using git ?

from somewhat 2007-2008 (I was learning Ruby then)

once you've gone through the pain of learning git, it's kind of unusual to move back to darcs

(now am fond of Elixir (and functional languages in general), but have to use python at work…)

do you prefer the simplicity for yourself, or for others you are collaborating with ?

I do all and anything myself alone (small projects)

I liked Darcs for it being functional and have a proper theory behind it

(also heard of Pijul, tried it, but feel its not ready yet)

Right, it’s not

well, welcome

I'm rather new to simplicity at all (3 months on FreeBSD after 15 years on GNU/Linux)

:)

I’m hardly using darcs myself but I maintain hub.darcs.net, you might find it interesting

Now learn sr.ht, tcsh and like how simple things could be while still working

oh does sr.ht support darcs ?

Already: hub.darcs.net/korobkov :) :0

I very much want them to support Darcs, still not sure

cool

:)

Idecided to rewrite all my school/university works on Elixir ;)

Was on Github and Gitlab before, but became fond of simple things. Hope not too late (hope they will still exist out there)

more people are learning about simplicity every day ! :)

:) My newest simple things are gemini:// and gopher:// via Elpher (I really like EMACS :))

I was very impressed how the Internet could work without JS and uBlock-like filters

I even liked the idea of creating my own site in gemini:// space (very simple markup), but have nothing to say to the world…

You can always start with “how I made this blog” :)

I've also read about your (Darcs) way of doing CI (darcs.net/Development/BuildBot)

I haven’t heard of Gemini:

while being a web programmer, I still don't have any site/blog :)

Gemini is very, very new really

I will just warn you, a lot of folks still love darcs but many docs are out of date

(at least I could have a simple personal page)

there’s no buildbot now

(Afk a little. See you o/)

thank you for the warning, I feel sad seeing at bugtracker with lots of open bugs and near no activity…

OK

I will read more about creating blogs, but are you really sure every person could/should have one? Isn't it more than enough information on the net already? What unique things should one have to start a blog?

About Gemini: gemini.circumlunar.space

No buildbot? I hoped it could be some alternative to sr.ht… Sad… :(

It all looks like I came too late… But I still hope it wouldn't go away completely! More people coming with simple things mindset may fix old issues, or even create something new (like Pijul)

About git: I didn't find it too much difficult to learn/use (at least for my own simple cases, and especially with Magit (EMACS git helper)), but being forced to use python while still being fallen in love with Elixir made me think a lot about harm of the monocultures…

When I was writing in Ruby, there were no other ruby programmers around in our university. But now Elixir is seeming even more experimental, and I just wasn't adviced/allowed to continue with it. While I myself see any functional languages being very niche and not having very mature library ecosystems, I didn't insist on starting next project in Elixir.

After such lesson, I started to look around in search of rarely used technologies to find if they are still shine, even abandoned by majority

About hub.darcs.net: I always was a Free Software developer (all my code, except drafts is free), and while having accounts first at github, then at gitlab, I thought my code would be of some use by people (clones, forks, issues), but as my projects were too small and somewhat specific to our university, seems no one else actually do/will use them. So, Darcs hub is in no way less good for me, than gitlab, while being much more

simple. :)

re amending a depended-on patch, yes, use rebase. "rebase suspend" to suspend the depending patch, amend the depended-on patch, "rebase unsuspend" to get back the suspended patch

thank you, Heffalump, I hope it should help, but my task in general is to amend the whole repo at once (like in `git-filter-branch`) to be able to import my old git repos, containing cyrillic patch names

korobkov: if you're just doing an import why do you need to amend old patches?

while having cyrillic patch names works well with newly created darcs repos, there still seems to be some encoding issues while importing fromgit

oh, right

well, you could rebase suspend the entire repo then unsuspend it

I could try it, OK

FYI there is a reasonable amount of activity on the darcs-devel mailing list (and collaterally on some of the active tickets) though only really between two of us (me and Ben Franksen)

glad to hear it :) So there is some hope for me for improving import on Darcs side?

(export fully OK with cyrillic)

thank you for mentioning it, Ive just requested subscription to the darcs-devel

while not knowing Haskell at all (just a litlle of Elixir) I may possibly help with some testing things

(as I plan to use Darcs as my promary and only VCS and Darcs hub as my priary and only public repository)

are you using darcs binaries or did you build it yourself? I think Ben did make some fixes to convert in the 'screened' repository (i.e. the latest development branch, not fully reviewed)

It also broke the test on Windows which will still need to investigate, but hopefully it did make things better on non-Windows :-)

I like to build everything from source myself! :) And I already built everything myself (FreeBSD). It's like baking my own bread (yes, I really bake my own bread :))

sorry, I didn't have/use any Windows since 2006 :)

(neither at home, nor at work, nor anywhere else)

so if you grab this repo and build it, you'll have the latest development code: darcs.net/screened

slightly more stable is darcs.net/reviewed

I already have GHC (as part of building Darcs port), but am new to Haskell. Still want to build it myself :)

would read manuals about it, OK

if you managed to build darcs from the released version, you should be able to build one of those branches quite easily. Just have a look and see if your import issue is already fixed.

and, if it's not, maybe you could write a test case that shows what's wrong (it would go in tests/failing_XXX)

Never thought of my acquaintance with Haskell would be like this :) Still no panic, just need some time to sleep! :)

:-)

and how would my custom build coexist with system darcs?

(or should I just make a special FreeBSD jail for experiments with Haskell stack)?

except if you use rebase, the on-disk repo format will be compatible

so it'll co-exist pretty safely, just use whichever binary you want at any given time

OK, thank you

if you do a rebase then I'd stick with the production version for now, because we are messing around with the format (particularly in 'screened').

FWIW personally I use cabal rather than stack, so can better help with problems with that

though AFAIK stack also works fin

e

so no need in building the whole stack?

if you already have cabal then nope

that has fewer dependencies to build from source

(of course with a bootstrapped language like Haskell you will always be depending on a binary _somewhere_ :-) )

still deosn't have, but should build likely without issues

:) It took some hours to build GHC :) (not enough RAM)

OK, % cabal --version

cabal-install version 2.4.0.0

compiled using version 2.4.0.1 of the Cabal library

so in a darcs source tree, cabal v2-build darcs should get you a darcs binary (or cabal v2-install darcs to also copy it to your .cabal/bin folder)

like to hve some compiling ahead (While was on Gentoo I liked to sit and look at GCC output… :))

on FreeBSD quarterly there not a lot of things to compile (while on ports tree head still too much to compile every day :))

% ghc --version

The Glorious Glasgow Haskell Compilation System, version 8.6.5

would be OK?

% nice cabal v2-install darcs

cabal: Ambiguous target 'darcs'. It could be:

exe:darcs (component)

lib:darcs (component)

exe, likely?

(or both)?

exe:darcs

yes, 8.6.5 is good, it's the one I use

I think

Yes, it's building:

darcs-2.15.2 and shelly-1.7.1 :)

yeah, we ended up (hopefully) temporarily inlining shelly into the darcs code because upstream kept breaking

so, screened is building, while reviewed is still copying patches

that's a bit surprising, because they get cached, and reviewed is a subset of screened

so I'd expect it to find the cached ones now rather than re-downloading them

should I check my darcs cache config?

what am I to provide?

it should happen by default - you could check the contents of ~/.darcs/cache - but maybe it's not worth worrying about too much

you'll get reviewed downloaded in the end :-)

I'm never in a hurry (while still have some hope to have supper while my server is compiling the Darcs ;)

done with dependencies, now the darcs itself

seems strange that Haskell utilize only one CPU core (while Erlang/Elixir use all in parallel..

might be a bad default - -j should cause a parallel build

so, the language itself is fully thread-safe, isnt it?

yep, and has a parallel runtime

glad to know :)

… done

% ~/.cabal/bin/darcs --version

2.15.2 (unknown)

great

lets test with my issue…

no, it's still there :(

* Ð<U+009E>бновленÑ<U+008B> gem-Ñ<U+008B>

:-(

but now I've compiled custom darcs first time :)

Feel free to raise a new issue, or comment on an existing one. While the issue tracker is a bit of a mess of old unloved issues, there is at least some chance it'll get looked at

OK, I'll try this

encoding is a bit of a disaster area, for reasons I don't fully understand :-/

(in the darcs code that is)

in Darcs, or in Haskell in general?

where in the code should I look first (on import)?

in darcs

umm, Darcs.UI.Commands.Convert I think

src/Darcs/UI/Commands/Convert/Import.hs I guess

what should I know about encodings in darcs (how things supposed to work)?

that's where I get a bit confused.

(I meant some specs or other documentation on encodings)

(or the code itself is self-speaking enough)?

well, I'm not sure there's any more authoritative source, let me put it that way

From bugtracker I've heard of some difference in handling data in patch names and content

is it on purpose?

(I meant, is it something that may/should be fixed)?

by content you mean the actual diffs, right? I guess you would never want those to get translated.

yes

Whereas patch names you want to display helpfully to the user, so there are cases where you want to re-encode them.

I _think_ the overall idea is that they should be stored as UTF-8.

I just thought that patchnames may also live without any intervention from Darcs…

aren't they already storead as UTF-8?

yes, but that requires intervention from darcs, if you as the person recording them isn't using UTF-8.

But clearly nowadays most people will be using UTF-8, and something is going wrong for you.

but this is where I don't want to claim too much about what should have happened for fear of getting it wrong

maybe in some future versions there wouldn't be any need for support of encodings, other than UTF-8?

I think if it was being written from scratch now, there would have been a stronger assumption around that.

or may be I could just compile my custom version of Darcs, just for importing, as export is already OK

But what we have now has evolved over time leading to the messier situation.

may be just some simple workaround?

I'm not quite sure what you'd change though.

I understand; at that time not all was good with encodings…

I'm trying to understand, at which point/stage exported UTF-8 stream from git is being corrupted/reencoded/stored in non-UTF way

I assume your own shell/environment is UTF-8?

could it be possibly stored as-is

yes, for sure

just having a quick look through the code

fully UTF-8 (since that old 2007-2008 years of that old repos)

that repos were created with the same LANG=ru_RU.UTF-8

makeinfo author message tag = do

let (name, log) = case BC.unpack message of

"" -> ("Unnamed patch", [])

msg -> (head &&& tail) . lines $ msg

(author'', date'') = span (/='>') $ BC.unpack author

the BC.unpack in that code looks highly suspicious

as do all the Char8 imports at the top of the file

doing that basically screams "screw up non-ASCII characters" :-)

because those modules are just about wrongly converting a stream of bytes into a stream of characters by just embedding them as 8-bit patterns.

(or vice-versa by truncating)

so, you already know what's happening? :)

well, sort of. As I said, it looks suspicious. I'm not quite sure what the simplest fix would be.

also, export is using Char8 heavily too and you say that's not broken

let me try it once more…

% d l

patch 6cb43d23fd31c49bdadd716e68e321e17d53ae89

Author: Andrey Korobkov <alster⊙vs>

Date: Wed Jun 10 01:51:27 +04 2020

* Тест UTF-8

this is darcs log

from a newly created repo

and now export…

% darcs convert export

progress (reading repository)

progress 1: Тест UTF-8

commit refs/heads/master

mark :1

committer Andrey Korobkov <alster⊙vs> 1591739487 +0000

data 14

Тест UTF-8

M 100644 inline test.hs

data 8

Тест

progress (patches converted)

progress (cleaning up)

progress done

all seems pretty OK, should I also test this stream on `git fast-import` ?

I'm just making a quick patch to test my theory, if it builds locally you could give it a go.

% git log

commit a837ff723e02be8a0d3d89134ec2329ec54a1e67 (HEAD -> master)

Author: Andrey Korobkov <alster⊙vs>

Date: Tue Jun 9 21:51:27 2020 +0000

 

Тест UTF-8

Absolutely OK!

(including patch data)

(this all is on system darcs 2.14.2)

Thank you in advance! :) I would like to build and try it myself too…

due to some really annoying problems with Windows and my corporate remote access software getting in the way, it took me longer to upload that somewhere than it did to write and build it locally

note that I haven't actually run it at all, but it does type-check :-)

:) :) Let me clone and build it…

got this:

% d clone urchin.earth.li/darcs/ganesh/temp/darcs-import-utf8-hack

Copying patches, to get lazy repository hit ctrl-C...

Exception while getting patches pack:

user error (Couldn't fetch `0000006835-3972772fbe917c5eb4d79be799b48f34ca18dc39bb01b958a1789041a2d7b4a2'

in subdir patches from sources:

 

thisrepo:/usr/home/korobkov/work/devel/darcs/darcs-import-utf8-hack

cache:/home/korobkov/.cache/darcs

repo:.

repo:https://urchin.earth.li/darcs/ganesh/temp/darcs-import-utf8-hack

)

bah

 

darcs failed: Couldn't fetch `0000006835-3972772fbe917c5eb4d79be799b48f34ca18dc39bb01b958a1789041a2d7b4a2'

in subdir patches from sources:

 

thisrepo:/usr/home/korobkov/work/devel/darcs/darcs-import-utf8-hack

does the new repo exist now? if so go into it and do a darcs pull

cache:/home/korobkov/.cache/darcs

repo:.

repo:https://urchin.earth.li/darcs/ganesh/temp/darcs-import-utf8-hack

oh, wait, the permissions may be screwed up

OK

I'll retry when you say it's ready

I'm able to see it in browser

file content is also accessible

ok, I fixed the permissions. It was that one specific patch. Try the 'darcs pull' from inside the repo you cloned already.

no, not just that one specific patch, hangon

its not cloned for me… :(

right, now really try

ok, just redo the clone then

OK

Copying patches, to get lazy repository hit ctrl-C...

seems OK now, thank you :)

korobkov: "I started to look around in search of rarely used technologies to find if they are still shine, even abandoned by majority" - I like this! And thanks for mentioning Gemini, it sounds fun

Finished cloning.

:) :) I glad to hear you like it. A also liked Gemini such to even decided to have my personal internet presence point (they cll it "Capsule" only in Gemini, not in WWW)

A good and helpful community is another bright side of some rare technologies. While mainstream techs have lots of documentation and forums and all that stuff, more simple techs could even not require such amount of knowledge at all :)

darcs is reputed to have an intuitive UI. I've been using it so long it's hard for me to really know if it is or not ;-)

:)

I suspect some of the flags could be more intuitive/ergonomic

For me to write classic (even not SPA) WWW application I should know not only DB and backend (which I do like), but lots and lots of frontend frameworks (I really don't, so just have to use pre-made components, or Django). In Gemini/Gopher space all could be much simpler.

and the are you sure prompts >:|

but I think it's intuitive overall, very much compared to git

no gemini server yet!

They think good server/client could be written by weekend in 1-2 hundreds LOC

rather impressive after firefox/chromium

About darcs and flags: I like the idea of preset flags via config ;)

git doesn't have it

that can be handy when you don't like the defaults, but can also create more failure modes and complicate troubleshooting

About git: Ive learned it before darcs and didn't find it too difficult for my own simple tasks, but only unless things need to go even little harder: I use only rebase workflow (without merges, so I actually needed Darcs from start, not git). And when rebasing the same patches once and once again it always went south.

Until I've read the problem is already known and is fixed by using patch-based VCS-es.

I really like git's command aliases

(and I really needed them!)

(and now my git commands are unique to me and not really shareable!)

never heard of them before (always thought it is done on OS shell level…)

my current difficulties with darcs is somewhat unevident patch dependencies

you can define them under [alias] in ~/.gitconfig. Super convenient, made git usable for me

(like here hub.darcs.net/korobkov/x0001/changes)

the UI tries to do the right thing with depenendencies implicitly, but I agree it can be confusing

there is some graphviz output to visualise them

[alias]

l = log --format='%C(yellow)%ad %Cred%h%Creset %s%C(bold blue)%d%Creset' --date=short -10

sort of thing.

sm[m]: thnak you for advices in git, but I really hope I wouldn't need it at all after fully switching all my works to darcs :)

my darcs equivalent has to be managed in the shell config:

re dependencies, it looks like hub's /patches shows them but /changes does not: hub.darcs.net/simon/darcsden/patches

Heffalump: I have absolutely nothing agains UI, my words were about darcs thinking of patches as dependent, even they have in common just neighboring lines (may be its so intended)

korobkov: ah, that's kind of necessary because of the theory

if two hunks touch (rather than overlap) then they cause a dependency unless both of them both add and remove lines

(I think that's the rule)

if both of them just add or both of them just remove, darcs wouldn't be able to put them back in the right order after a cherry pick

yes, and Elixir makes thing somewhat worse hub.darcs.net/korobkov/x0001/patch/…0f9d5705b90967785f3d8af112a54de9023

semantically the patch shouldnt be thought as dependent on previous one, but darcs marks is as such, just because of single comma! :)

ah yes. That's a general problem with VCS and lists though

I try to format my code so adding new elements at the end of things doesn't change the previous line.

can you use leading comma style ?

sm[m]: never thouth of it, as I usually like the defaults and just used to like them after some time

I dont think it's bad at all, just need some accuracy while merging

yeah. Some languages allow a trailing comma to help with this

it's too much to ask a general-purpose VCS to know if that comma is important, I think

so darcs has to assume it is

ruby have %W() syntax, Elixir also have sigills, but they arent useful for dependency declarations, as they are not always simple

another issue is positioning blank line before newly added method at the end of file

sorr, not able to illustrate with the code, ok

korobkov: did you manage to try that patch yet btw? (no worries if not, just wondering)

near forgotten about it, sorry :( was building all this time

had built, trying…

sorry, Fremantle монÑ<U+0082>иÑ<U+0080>Ñ<U+0083>еÑ<U+0082>Ñ<U+0081>Ñ<U+008F> на коÑ<U+0080>невой URI

hmph

still not got it… :(

git fast-export master | ~/.cabal/bin/darcs convert import <path>

I think your best bet would be to construct a shell test that tests for the 'correct' output, it'd make it easier for a developer to experiment with the code

~/.cabal/bin/darcs --version

2.15.2 (unknown)

OK, it would be simplier even for myself to test

what should I return in case of success/error?

what are input parameters?

fail (i.e. non-zero exit code) for error, success (zero exit code) for success :-)

as ususal :)

look at one of the scripts in tests/ for an example, they all follow a similar pattern

you could just attach it to the bug report, or name it failing-issueXXXX-convert-import-utf8.sh or something like that and send in a patch with some name like "accept issueXXXX: convert import double-encodes Russian characters"

what shell should I use? bash-compatible?

yes, bash (if you look at our scripts they all start with #!/usr/bin/env bash

thank you for your work, so now we are sure they are double-encoded

sorry, I'm too sleepy now :)

of course it's bash :)

so, I could even send it by email and it would work? Never had sent one by email, but many times heard it was the main way of sending patches some time ago. Should it work now?

yep, though darcs now defaults to not doing it because so few people have sendmail configured

but if you have it configured ok, then darcs send --mail in a clone taken from screened will automatically send the patch to our tracker

is it so difficult to config it such a way?

where could I read about it? (it's one of the things I like darcs for)

or darcs send --help

OK, I'll try it to learn do things the ways they are expected to be done :)

(but my Gnus in EMACS is also very good)

you can also just let it save a file, and attach the file to the patch tracker directly

yes, of course, but sometimes I still use majordomo to subscribe to mailing lists :)

:-)

:)

I stopped running majordomo a few years back. I liked it, but it was just a bit too clunky and hard to manage compared to mailman.

things were working even before www…

that's another reason for me to still use some old techs, as if no one would, then they could just disapper, and world would became more sad place (`man monoculture`)

and once again, seems like I've came too late for lot of interesting thins of the past

well, if I could write proper test for cyrillic import it could help some other people to continue to use darcs :)

(and me counting first)

and what would be the future of the hub.darcs.net? Isn't it too difficult to run? Won't it dissappear?

sm[m] runs it. I think if he stopped I'd try to step up. But as with most of the darcs setup we don't have the critical mass to guarantee anything forever.

korobkov: it's not too difficult so far. But I only provide "essential maintenance" for a few years now

A'm not at all about any guarantees :) Just to be sure you are not tired with it enough to abandon it all…

when darcs gets more active, darcs hub gets more interesting

no, don't worry it's not in danger, and if I stop I'll ensure a handover

one of these days some way forward will open up

sm[m]: glad to hear

trying to get the rest of my projects/life in order, maybe that will help :)

it would be too sad to lose darcs hub

I would like to spend the time to find out what's happening on darcs hub :)

and set up a way to communicate with users

aren't mailing list and irc channel not enough?

darcs hub users aren't the same

well

(I don't know how much the overlap is)

last time, I exported all email prefs into a mailchimp list, which was super manual

and has to be redone each time

of course I wouldn't wish you more manual work just to support old techs

I'm hee not to change the way things are going. I've just found nice, simple and useful tool and supporting repostory hub to use for all my projects, and want to fix some issues, which prevent me from use it for one and every of my projects. ^)

sm[m]: I also don't think or say all new techs are all that bad. I was able to open your matrix link in my eww browser in EMACS :)

As long as the tools are federated, not centralised, I could make use of them (while I strongly prefer the ones, which could be run from the EMACS «operating system») :)

sm[m]: , Heffalump: thank you all for your time and your work to help me. Now I really need some food and rest.

See you later, and I will learn how to write tests for darcs to make one for cyrillic import.

np, night korobkov !

:)