Hi Matěj,

Very belatedly, thank you for this patchset!  It's great to have a
more reliable method than email, and to have labels applied and long
messages split automatically.  And to have tests.

I've applied this now.  I did some minor follow-up work, mostly to
make the code more strict, but also to restore label caching and fix a
corner case of using --mode=email with split messages together with
skipped confidential tickets.  (I skipped the "Bit reformatting"
commit, minor personal preference for the existing state of things
there.)

So, thanks again.  Can we add a copyright line for you?  You've done

On Sat, 01 Jun 2024 23:44:04 +0200
Matěj Cepl <mcepl@cepl.eu> wrote:

> On Sat Jun 1, 2024 at 5:57 PM CEST, Bryan Gardiner wrote:
> > I saw the mismatch between #223 and #225 on your tracker, and I'm
> > trying to get a complete run of m2crypto myself using hut.  One
> > issue was that it was trying to create labels multiple times, so
> > locally I've removed @cache from get_labels().  
> 
> Yes, you are right: premature optimization is the root of all evil.

One of the two hard problems in Computer Science...

> > Is it possible that you did the run with the #223/#225 issue before

On Tue, 21 May 2024 08:04:16 +0200
Matěj Cepl <mcepl@cepl.eu> wrote:

> On Tue May 21, 2024 at 5:38 AM CEST, Bryan Gardiner wrote:
> > This has me wondering how much work it is to talk to Gitlab’s API.
> > Maybe that wouldn’t be too hard to hook up.  Although then we would
> > have to drop ‘lazy’ from the name...  
> 
> My first priority is to get M2Crypto issue tracker up and going,
> not to get ideal solution. So, I will certainly work out of the
> downloaded files. Switching to GitLab API will be certainly left
> as an excercise for the reader.

For sure.  I didn't mean for this to be a request.  I might try it,

Hey Matěj,

On Thu, 23 May 2024 14:38:01 +0000
~mcepl <mcepl@git.sr.ht> wrote:

> So, close and yet so far!
> 
> It almost works, https://todo.sr.ht/~mcepl/m2crypto/ is something
> which really resembles the useable thing, but it isn’t. Could you
> tell me why https://todo.sr.ht/~mcepl/m2crypto/223 is really not the
> equivalent of https://gitlab.com/m2crypto/m2crypto/-/issues/225,
> although it claims it to be so?

> Note that I have `--create-missing-issues` there, so supposedly

On Mon, 20 May 2024 00:18:41 +0200
Matěj Cepl <mcepl@cepl.eu> wrote:

> The following changes since commit 350d5fd31d5255b71041858c6a2085d3b1d8daa8:
> 
>   Add git config for subjectPrefix to the readme. (2024-05-06 23:21:48 -0700)
> 
> are available in the Git repository at:
> 
>   https://git.cepl.eu/cgit/git/lazygl2srht 
> 
> for you to fetch changes up to 66d4ce7856cf5d54689b5a606d0f4b1d2cf2d3dc:
> 
>   feat: check existence of labels (2024-05-20 00:13:30 +0200)

>>>> On Sun,  5 May 2024 08:52:15 +0200
>>>> Matěj Cepl <mcepl@cepl.eu> wrote:

> Ordinary print() function in Python is buffered (there is flush
> option, but it alleviates just part of problems). Logging module
> knows about and deals with the problem.
> 
> https://stackabuse.com/bytes/flush-the-output-of-the-print-function-in-python/
> https://stackoverflow.com/q/230751/164233
> https://stackoverflow.com/q/55619733/164233
> 
> (logging.INFO changed during the development of the code to
> logging.DEBUG as needed).
> 

>>>> On Sun,  5 May 2024 20:41:27 +0200
>>>> Matěj Cepl <mcepl@cepl.eu> wrote:

> Formatting produced by Black is commonly agreed one, and it
> simplifies life a lot.
> 
> Add black to your pre-commit hook.
> 
> Signed-off-by: Matěj Cepl <mcepl@cepl.eu>
> ---
> I was trying to be conservative (particularly changing all ' into
> " makes huge diffs), but if you don't care, I can go all the way
> in.
> 

On Sun,  5 May 2024 09:35:06 +0200
Matěj Cepl <mcepl@cepl.eu> wrote:

> Formatting produced by Black is commonly agreed one, and it
> simplifies life a lot.
> 
> Add black to your pre-commit hook.
> 
> Signed-off-by: Matěj Cepl <mcepl@cepl.eu>
> ---
>  import_issues.py | 306 +++++++++++++++++++++++++++--------------------
>  pyproject.toml   |   3 +
>  2 files changed, 182 insertions(+), 127 deletions(-)
>  create mode 100644 pyproject.toml

On Sat, 04 May 2024 18:12:18 +0200
Matěj Cepl <matej@ceplovi.cz> wrote:

> On Sat May 4, 2024 at 9:26 AM CEST, Matěj Cepl wrote:
> > 1. STARTTLS support (that’s that patch)
> > 2. Some kind of logging … to have a script running for four hours
> >    ((3000*5)/3600) without absolutely anything on stderr is
> >    rather cruel.  

More logging sounds good, but can you see why the initial startup is
taking so long?  Is it slow for --mode=print too, or just some SMTP
slowness with --mode=send?

> One more observation from running M2Crypto attempted conversion

On Sat,  4 May 2024 09:32:53 +0200
Matěj Cepl <mcepl@cepl.eu> wrote:

> Signed-off-by: Matěj Cepl <mcepl@cepl.eu>
> ---
>  import_issues.py | 13 ++++++++++++-
>  1 file changed, 12 insertions(+), 1 deletion(-)

Applied, thank you!

- Bryan