Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 23 May 1995 00:16:48 +0200
From:      Julian Howard Stacey <jhs@vector.eikon.e-technik.tu-muenchen.de>
To:        asami@cs.berkeley.edu (Satoshi Asami | =?ISO-2022-JP?B?GyRCQHUbKEI=?= =?ISO-2022-JP?B?GyRCOCsbKEIgGyRCOC0bKEI=?=)
Cc:        jhs@freefall.cdrom.com, CVS-commiters@freefall.cdrom.com, cvs-ports@freefall.cdrom.com
Subject:   Re: cvs commit: ports/comms/hylafax - Imported sources 
Message-ID:  <199505222216.AAA07935@vector.eikon.e-technik.tu-muenchen.de>
In-Reply-To: Your message of "Mon, 22 May 1995 03:42:50 %2B0200." <199505220142.SAA18778@silvia.HIP.Berkeley.EDU> 

next in thread | previous in thread | raw e-mail | index | archive | help

Satoshi,

You've made several references to my supposedly rotten Hylafax port ...
- You wrote I had comitted a bad port, 
- You threatened to delete it in 24 hours.	 !
- You hacked it & the patches don't apply to the new (wrong) distfile,
  (& to be fair, some bits are improved, thanks).

I agree what now lies in ports (Ref. ctm 438) does Not work, that is because
I made one mistake, & you made more :-(

Your threat/promise to delete it, I treat as mixed, as I don't want that
hacked up port in there, but I do want my working port (& yes, I'd like 
your good hacks on top, too please :-)

Fortunately I've realised where we both went wrong (in different places,
useful to know to avoid for future)
& offer 2 alternative solutions further below	:-)

----

First however, let me disprove your statements about a bad port:
Yes FreeBSD Hylafax is broken !  It is also Not what I committed.
Anyone who doubts My version Compiles & Works: I can prove it - just mail me :
	- Your email address for private email receipt of a 70K typescript
	  of hylafax (my version as originally committed, recompiled today
	  on my host (commands from ctm src 645, older kernel)
	- Your fax number, so I can fax you (or preferably a friend you
		name in Germany (cheaper for me than faxing internationally) !
	- Your login name & a some way encrypted password for this host,
	  so I can grant your temporary login here when I'm next dial-up slip
	  connected.
There are people on this list that know I have a working Hylafax here.
I've already faxed them with it weeks ago.
Anyone who merely wants a port or package of Hylafax on FreeBSD can also
`roll their own': grab the distfile I worked from, & my original port tree in 
	freefall.cdrom.com:~jhs/hylafax*
(PS my home system is currently based on src ctm 645, except older kernel).

----

> All I did at that time was to make some comments on your 100+ line Makefile,

Useful comments too, most (perhaps all) incorporated, Thanks :-)

----

> Julian, I asked you to send me a mail when you are done, not to commit it.

That puzzled me, it contradicted what you had posted earlier,
The attempted interactive micro-management of this port initial commit was
unexpected, you shifted targets from a `Hylafax now for the CD',
to a stance of `ideal styelised port of Hylafax, no matter if it makes it
on to the CD'.

You had earlier written you wanted hylafax ( & had chased for it).
You had earlier written we were in code freeze _except for new ports_ 
	(go check your outgoing posting log if in doubt)
So I worked hard & long, exclusively on Hylafax, to get a working error free
port to contribute, then as my IP access availability was intermittent,
& as talk sessions between us don't work well, what with timezones, & tiredness
one end, & a ticking stopwatch reminding me of telecom modem charges the 
other end, I commited the New Port while I could, as soon as complete,
to give us a Working Port before the impending CD deadline.

Interactive work between us isn't ideal, eg when you're tiredly composing a
`talk' message & I have an appointment pressing, while also hearing my
stopwatch ticking at peak telecom modem rate.
I'm not on line except when I pay to be, & it is not a good idea for us to 
try to work interactively together on one port. Better either of us alone.
Interactive is better in adjacent timezones, with permanent IP connectivity,
not dial-up.
I understood I had a mandate to provide a New Port,
I considered myself bound not to change it after commit, & to consider it
frozen, _once_ it was integrated, but I had not realise you intended to
micro-manage this little project prior to initial commit.

----

> I haven't taken a look at any of your numerous
> auxiliary scripts and I haven't done any testing.

Don't worry, `cd /usr/ports; make' does NOT use this tree yet
	(I very deliberately added no SUBDIR += hylafax in comms/Makefile), 
thus it Cannot break the ports tree, precisely until such time as you have had 
time to look, & do decide to add SUBDIR += hylafax.
Unless you own & rely on a fax modem as I do, my actual software_in_use 
as compiled, installed, & used on day of commit, is more `tested' anyway.

----

One thing I did not do was refetch the distfile, to check it had
unchanged content, (I didn't as I have a _slow_ link here in Munich),
I _did_ check it was still on the same distribution site with the same name.
I did not ftp it to freefall/thud & run md5 or make on it there, as that would 
have been difficult & slow (it's 30M), though I could have nohup make'd it.
Anyway, that's what files/md5 checks for, & even if I'd checked it, it could
have changed straight after (& indeed may change yet again (hope not)).

----

When you took my commited port,
You ignored my original committed md5 (ab3bdeee68a6cd64c270e09dffe928e7)
& commited a new md5 (014ec8a64ea7493d26c50c61b5ef3111).
You applied my patches, which failed, 
you seemed to assume I'd written bad patches too !
If an md5 complains, & patches blow: consider the possibility the distfile
may have changed ! Don't just commit a new md5.

-----

> You haven't demonstrated you can maintain the port (see below),

Totally Unfair !  You denied me any chance to !
In the ~14 hours between commiting this & then getting the mail, first I was
asleep, then I was waiting for my mail link to restore itself.
You broke Hylafax by changing md5, & now accuse me of bad maintenance !

----

> this isn't an excuse to import something I can't understand.

I'm not looking to excuse myself for something you don't understand.
Don't touch what what ain't broke !

----

> I do NOT want a broken or messy port in the tree.

No one wants broken ports.
You imply a messy port that works is not wanted ?  
Tell that to the folks that want fax capability !
At least what I commited _works_, which is more than some ports/Makefiles do.
Anyway first `messiness' was a workaround the broken handling of DISTNAME
in your area (share/mk/bsd.port.mk)
I'm glad to hear you've added EXTRACT_SUFX, I've had a quick look, 
I'll look more, when I get all the src ctms (IP has been intermittent last
2 days, not all ctms here).

----

> ... BROKEN list, ... 

I never raised this topic re hylafax, you'r confusing another thread.

----

> If you can't fix this within 24 hours, expect it to be purged from the 
> repository and replaced by another, simpler and working version.

RIDICULOUS, - YOU IGNORED THE WRONG DISTFILE MD5 WARNING
Tossing out the port after you've micro-managed it & it's now screwed up would
not be a good example to encourage new people to start contributing ports.

2 SOLUTIONS I SEE
	A (Quick Reliable)
	- Back Out (& save as diff -c) all your changes.
	- Copy the generic distrib file in freefall:~jhs 
	  (that I got from sgi.com), to some ftp'able place on the Internet,
	- Change DISTNAME & MASTER_SITES to new location.
	- Optionally selectively reapply your changes.
	B)
	- Fix the current contents of ports.

I'd prefer (A) I'm not keen to do (B), but
have no further time for either for a few days, 
Real Life (TM) now calls.

I'd be happiest if you did (A) prior to CD-pressing on freefall, 
& you will get back a working port with the supplied distfile.
(B) is less atractive, (& who knows if distfile will change again).

Julian Stacey	<jhs@freebsd.org>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199505222216.AAA07935>