IRC log FDSCo meeting 10-Oct-2006

[Karsten Wade]

You can skip over the overly long attempt for me to grasp PO, POT, and
other dynamics ... :)

<quaid> <meeting>
<quaid> greets, amigos
<stickster> hola
<megacoder> hi
<quaid>
http://fedoraproject.org/wiki/DocsProject/SteeringCommittee/TaskSchedule
<quaid> and such ...
* stickster notes Bob might miss this one, understandably... sends best
wishes
<quaid> let's capture AOB, too
<quaid> BobJensen-Away++ good luck
<quaid> <aob>
<quaid> * quaid's proposal of merging Web-only update and new
translations
<quaid> * quaid to GSoC Mentors Summit, chance to get in a word edgewise
<quaid> other's?
<quaid> others'?
<stickster> errr.... nope
<quaid> ok, just taking a chance to make sure we have time for all
* stickster flips users, back in 60 sec
<quaid> ok, AOB is still open to finish whenever
<quaid> screen++
<quaid> "More than one-third of U.S. college campuses have Wi-Fi, giving
their students blanket wireless broadband access at school, according to
data released today by the non-profit Campus Computing Project."
<stickster> back
<stickster> Next, the airports... then if Starbucks frees wifi, that
pretty much takes care of the rest of CONUS
<megacoder> I'd like to see a blanket with WiFi access...
<megacoder> and maybe a USB port.
* stickster hopes quaid edits this out of the meeting
<quaid> why?
<quaid> I never do :)
<quaid> ok, continuing on ..
<quaid> Paul: anything you need for the IG?  
<stickster> Yes, I need trans to dump the POT into their system
<stickster> So the translators can work on it!
<stickster> Maybe I need to investigate getting us some collateral
access
<stickster> I did send the requisite email to f-trans-l to the usual
response of utterly deafening silence
<quaid> hmm
<stickster> But, to be fair, haven't been able to check email since
yesterday
<stickster> Will report in a min.
<quaid> crap-O!
* quaid thinks
<quaid> it's highly likely I only need to be added to the CVSROOT/avail
<quaid> I think I have an account on that machine
<stickster> Like spaghetti-O's only a little less popular
<quaid> but from when I asked how to do it, it was a bit of spaghetti to
add a module
<quaid> now, if it is just 'cvs ci foo.pot' then perhaps I can do it ...
<quaid> do we know what the cvs path is?
<stickster> I think it ends up being a msgmerge in between
<stickster> /usr/local/CVS  IIRC
<quaid> alam: help!
<quaid> noriko: is snowlet around?  we could use some help with a "Very
Important" Fedora Translation item :)
<noriko> quaid:  I can not see him on any other channels
<quaid> noriko: thanks ... I forget if you are both in the same office
(sometimes)?
<noriko> yup
<noriko> the office is huge ;P
<noriko> I will look arround for you
<quaid> :), thanks
<quaid> ooh!
--> snowlet (i=snowlet at nat/redhat/x-609c80959c37ec1b) has joined
#fedora-docs
<noriko> he is coming
<quaid> stickster: I am pulling down what looks like the stuff
<stickster> quaid: Excellent
<quaid> so let's ask snowlet 
* stickster thinkis wiki-izing this would be good
<quaid> if it is as "easy" as 'cvs ci -m "this file has changed on the
following lines for these reasons" po/fedora-install-guide.pot
<quaid>  echo $CVSROOT
<quaid> :ext:kwade at elvis.redhat.com:/usr/local/CVS/
<stickster> I thought it might need a "msgmerge -o new.pot old-stuff"
<quaid> stickster: did you catch the discussion on f-trans-l about the
proper way to treat the translators?
<quaid> ah! that might be it
* quaid knows not
<quaid> snowlet: we are trying to get the final POT file for the Fedora
Installation Guide into CVS.
<quaid> snowlet: I seem to have write access, so I could do it, which
would save a step for you or alam ... 
<quaid> but I really don't know what I'm doing
<quaid> :)
<snowlet> quaid, do you have the latest .pot file on hand?
<quaid> stickster does
<stickster> It's in our CVS too
<stickster> i.e. Everything's committed in Docs CVS, but the new POT
needs to go into elvis
<stickster> Wow, that almost sounded like he was alive again
<snowlet> quaid, you can copy the .pot file into your local machine, cvs
up to get the latest .po files, run "msgmerge" for all languages
(generally I'll write a script), then commit
* stickster confirms that yes, all IG stuff is committed
<stickster> snowlet, quaid: We have all that in our Makefile already
thanks to megacoder
* megacoder nods gratiously
<stickster> If you do "make pot ; make po" it should automagically work
* quaid tries to imagine this across the different $CVSROOTs
<quaid> sorry, am I committing ... updating PO files?
<stickster> quaid: Well, on our side, you shouldn't have to do a thing.
I've done all that for you.
<quaid> so ... I do make po ; make pot in the
cvs.fp.o:/cvs/docs/install-guide/ and then ... cp po/*.po to my checkout
from elvis?
<quaid> ok, so then just the second part?
<stickster> Yeah, I believe so
<stickster> (the 2nd part)
<quaid> cp install-guide/po/*.po /path/to/elvis/install-guide/po/
<stickster> That's the part I'm wondering about
<stickster> Should it be a "cp", or a "msgmerge"?
<quaid> or should I take what is in elvis and cp -that- to what is in
the install-guide
<stickster> snowlet: It doesn't matter to the translator if we get rid
of some of the strings in a .PO file, right?
<stickster> snowlet: Because they have a program cache on their local
system?
<quaid> and then re-run the make po, make pot
<stickster> quaid: Nope, definitely not that last
<quaid> k
<megacoder> The "cp" approach is correct.
<snowlet> stickster, no.  after translators do "cvs up", they'll lose
what they had
<megacoder> msgmerge is to insert the persistent (pot) into the
transient (po).
<stickster> megacoder, snowlet: thanks
<megacoder> stickster, you know this, you just confused yourself in the
frenetic pace
<stickster> So quaid can just cp our POT up, right?
<stickster> Or must the translators have an updated PO as well?  Are
they expected to msgmerge?
<megacoder> Yes.  The POT is plenipotential and stand alone.
<megacoder> Translators produce update PO which get merged into the POT
<stickster> Well there you go then
<megacoder> There is no carry-over info from PO v1 to PO v2; it's all in
the POT
<stickster> A neat system, but one does wonder why a translated string
isn't stored somewhere (say, in ~/.prog_name)
<stickster> It could be indexed with strfile and kept around for later
use
<stickster> OTOH I suppose that it's hard to come by a duplicate <para>
for example
<megacoder> You're welcome to do it that way, but the translators want a
POT with strings to be translated and placed into the PO which get
merged back...
* quaid had to do a kid grab, is back
<quaid> ok, so ... 'cp' then 'cvs ci'
<quaid> right?
<stickster> yup
<stickster> right megacoder?
<megacoder> cp new.po old.po; cvs commit
<megacoder> ci?
<quaid> is the same as commit
<megacoder> OK, I didn't bother to read that far.
<quaid> oh!
<quaid> ok, so I don't fool with the POT file
<stickster> Yes, you do
<stickster> megacoder is confusing you!
<stickster> megacoder: stop it!  :-D
* megacoder is not surprised, I'm confused myself
<stickster> cp docs-ig.pot i18n-ig.pot ; cvs ci
<stickster> But he's right, POT is just a non-translated PO
<stickster> I'm confused. Why are we here?
* megacoder sniffs
<megacoder> why is there air there?
<stickster> Oh yeah, the babes and beer
<stickster> D'OH!
<quaid> oh, because I asked about doing it instead of just asking
snowlet to do it
<stickster> he
<snowlet> msgmerge xxx.po xxx.pot > temp.po; mv temp.po xxx.po
<stickster> Oh yeah, that part too
<quaid> sorry, thought you said the Makefile did that?
<stickster> In our CVS, yes
<stickster> In i18n you have to provide it for the translators
* megacoder considers that there is no real reason to write this stuff
down in an FDP newsgroup
<quaid> sorry, my fault
<stickster> heh
<quaid> snowlet: can you please update the install-guide file from the
new POT/PO/foo file?
* quaid didn't realize so much time had passed
<quaid> I'm going to move us along
<snowlet> quaid, it the latest .pot on i18n.redhat.com yet?
<snowlet> s/it/is
<quaid> no
<quaid> maybe stickster can just email the bits to you
<stickster> snowlet: Can you get me access to do this?
<snowlet> quaid, send it to me at ccheng at redhat.com, please.  I'll put
it there, do all the merge and commit
<quaid> ok, I'm going to focus on getting a usable DUG draft out in one
week
<snowlet> stickster, no, sorry.  I am not the sysadmin...
<stickster> snowlet: Any idea to whom I should speak?
<snowlet> stickster, Matthew Schick (mschick) maybe?
<stickster> OK, will do
<quaid> eitch is not here to discuss trans/doc process, think he was
going to report next week anyway
<quaid> megacoder: any thinking or questions about the trans Web UI or
CVS migration?
<megacoder> I've been refreshing my python foo, but I think there is
little retouching needed for the UI to be repositioned.  There is a
mysql database that will need some admin support, but it looks do-able
so far.
<megacoder> More info next week.
<quaid> ok
<megacoder> I'm not sure what I need to think about CVS migration...
<quaid> yeah, I was just trying to spur FInfrastructure with more
resources :)
<quaid> ok, my AOB
<quaid> I was thinking, for fedora-release-notes ...
* megacoder dodges a bullet, methinks
<stickster> megacoder: Pretty slick, slick
<quaid> let's put up the ISO notes to go with the release date ...
* stickster notes Bob said same, +1 from me too
<quaid> I haven't seen much of substance in the the Wiki, so am
wondering if we need to bother to carry the content differences over
<stickster> Ah, that's an action item
<stickster> Need to find out from mmcgrath if he can remove the governor
from an account or two
<quaid> and was thinking that our Web-only update could be just all the
translations and minor bug fixes.
<stickster> Oh right, the tagging problem. Scratch my last comment. :-(
<quaid> right, see ... 
<megacoder> s/tagging/nagging/g
<quaid> we -could- do an update fairly easily, but it wouldn't be
translatable
<stickster> heh, that too
<stickster> It's that not-so-fresh feeling
<quaid> well, what do we think of this idea?
<quaid> have there not been much of substance changed in the Wiki?
--- tsNEET is now known as tsekine
<stickster> I think I'm following you, we could check each page's
info/log, and just manually insert the small changes
<quaid> well, or that :)
<stickster> oh
<quaid> no, that works
* stickster reads xml2po source to see where the tags are read
<quaid> I guess ... it's how many versions are we posting?
<quaid> 1. ISO (en_US + trans)
<stickster> ^^^ on the release date
<quaid> right!
<quaid> 2. f-r-n update package (sent to repo) == minor bugfixes +
finished ISO-matching translations
<stickster> + possibly working l10n'd menu entry
<quaid> 3. Wiki -> XML not trans
<stickster> Isn't #3 superfluous?
<quaid> well, back to that ...
<quaid> which was the question ...
<quaid> was there anything of substance in the Wiki to move to the XML?
<stickster> Oh, I meant, skip right to 4. Wiki -> XML -> translated ==
minor updates on everything from Step #2
<stickster> We might have material that's worth republishing, but if
we're going to bother, then we should have it translated and ship it all
at the same time
* stickster notes this is for Bob to weigh in on too of course
<stickster> If I can get mschick to give out some access that would help
a lot
<stickster> Translators won't mind redoing 15-20 strings, it's 800 at
once that's a problem :-)
<quaid> well, OK
<stickster> am I obtusely missing a big point?
<stickster> it's possible... hell, likely even
<quaid> one sec, half a brain
<megacoder> probably, but try to be specific
<quaid> ok, back
<quaid> I was thinking of letting folks finish what had been done
already rather than adding to it
<stickster> Oh, I thought that "half a brain" was you hurling an
epithet, not a description
<quaid> but if it is only a few dozen strings
<quaid> lol
* stickster backs up
<stickster> Ah I see
<quaid> it sure reads that way, don't it?
<stickster> jk
<stickster> I think I see your point -- I was thinking of a little
longer timeline.
<megacoder> Next week, I'm wondering what folk will be considering the
wiki experiment for inputting content
<megacoder> and how it's turned out.
* stickster hands megacoder an "I Hate Wiki" button
<quaid> stickster: well, I hadn't defined timeframe ...
<megacoder> But I'm not asking this week ;-)
<stickster> Right, right...
<stickster> ...let me make myself more clear: I think that...
<stickster> ...there are plenty of translators who have substantial
portions left to do.
* megacoder thinks that could be an "I Hate [A-Z][a-z0-9]+" button...
<stickster> We can set a target date for pushing a f-r-n update
post-spin.
<stickster> Work toward that with the current non-updated XML material.
<stickster> Also set a target date for pushing a second update with
updated material.  We could likely expect content returned from most if
not all of the translators who finished for the first update.
<stickster> I would think release + 7 days for the first update, then
release + 21 days for the second maybe...
<quaid> well, and that was the question ... did we want to not do a
content update for the package update?
<stickster> Not for the first, but for the second...
<megacoder> Hey, Glezos is reporting that the "adaptive index" ploy is
working.  Check the list.
<quaid> ok, so update the package twice and give plenty of time for
trans?
* stickster worries about precendent of letting translations slip out of
gear with each other in a single package
<quaid> that's another thing we can tweak in the update
<stickster> quaid: Exactly
<stickster> I will pursue mschick... I already sent the email, you and
Max are cc'd
<quaid> well, isn't first update just adding more translations?
<stickster> yes
<quaid> and second is new content + new content trans
<stickster> yeah baby
<quaid> I'm missing the slip out of gear part then
<stickster> it doesn't in that scheme
<quaid> ok
<stickster> I was just voicing some trepidation about having an update
of some sort that put en_US out of sync with the other translations
<stickster> megacoder: Cool
<stickster> megacoder: Unfortch, that's FF v2, which we won't be seeing
for some time
<megacoder> Yeah, but think how cool FC7 is going to be...
<stickster> right, sweetness!
<megacoder> We have the technology; we can adapt it.
<quaid> ok, this has gone on long enough
<quaid> anyone wants to discuss the GSoC stuff with let me know
<quaid> otherwise ...
* quaid prepares to close this meeting
<stickster> *whew*
<-- megacoder has quit ("I Worship His Shadow.")
<quaid> </meeting>
-- 
Karsten Wade, RHCE, 108 Editor    ^     Fedora Documentation Project 
 Sr. Developer Relations Mgr.     |  fedoraproject.org/wiki/DocsProject
   quaid.108.redhat.com           |          gpg key: AD0E0C41
////////////////////////////////// \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://listman.redhat.com/archives/fedora-dsco-list/attachments/20061010/ca288e06/attachment.sig>