What's the canonical way of reporting a bug against tcllib?
tomas
Hi,
I'd like to report a bug against tcllib (I'm not the discoverer, the
honour goes to John Roll [JBR], who describes the bug in
<http://wiki.tcl.tk/15512>. How do I do that?
It's a very small ommission: in line 116 of grammar_peg/peg_interp.tcl:
ict_match_token "Expected $ch"
should read
ict_match_token $ch "Expected $ch"
and yes, it's still in tcllib-1.13
<rant>
I tried mucking around in SourceForge, but after seeing all the red tape
I'd have to cut through (i mean: Job Title? Number of Employees? Hey, I
started entering bogus stuff, but then decided I'm too old for that).
And those are marked as REQUIRED fields.
Am I mis-interpreting SourceForge's interface? (very well possible!)
</rant>
So: is there any straight mailing list for that? I wouldn't mind having
to subscribe (spammers be thanked!).
Thankful for any pointers.
tomás
Arjen Markus
If you can not report a bug anonymously, you can always subscribe (I
guess without
a user account on SF) to tcllib...@lists.sourceforge.net
Regards,
Arjen
Larry W. Virden
> <rant>
> I tried mucking around in SourceForge, but after seeing all the red tape
> I'd have to cut through (i mean: Job Title? Number of Employees? Hey, I
> started entering bogus stuff, but then decided I'm too old for that).
>
> And those are marked as REQUIRED fields.
>
> Am I mis-interpreting SourceForge's interface? (very well possible!)
> </rant>
>
Yes and no. It does appear that SourceForge has dropped into some
kind of strange universe where everyone who wants to report an error
has to identify their job title and number of employees. Of course,
there's no way they can verify the information, so if you say you are
president of your own company, and that you have 1 employee... well,
can they really argue that idea?
I do agree, however, that is ridiculous.
There's something called OpenID, whereby you use an already
established login from Yahoo, Google, AOL (do they even still
exist??), LifeJournal, Flickr, Technorati, Wordpress, Blogger, Vidop,
Verisign, ClaimID, OpenID, or MyOpenId and use that to log into the
system.
If you happen to have one of those ids, then you should be able to go
the openid route.
Alas, too many spammers and hackers were taking advantage of the
openness of sourceforge (and other sites) to continue to allow
anonymous ticket openings.
Larry W. Virden
> If you can not report a bug anonymously, you can always subscribe (I
> guess without
> - Show quoted text -
I don't know if subscription through sf.net still is allowed without
an account. However, I believe that ActiveState makes a mirror of the
list available - but again, they probably require an account to be set
up.
It does not appear that the tcllib mailing list is mirrored into the
google groups environment.
However, bugs discussed on the mailing list do not get a ticket opened
on them - someone still has to log into sf.net to open the ticket.
Alexandre Ferrieux
>
> I'd like to report a bug against tcllib (I'm not the discoverer, the
> honour goes to John Roll [JBR], who describes the bug in
> <http://wiki.tcl.tk/15512>. How do I do that?
>
> It's a very small ommission: in line 116 of grammar_peg/peg_interp.tcl:
>
> ict_match_token "Expected $ch"
>
> should read
>
> ict_match_token $ch "Expected $ch"
>
> and yes, it's still in tcllib-1.13
>
> <rant>
> I tried mucking around in SourceForge, but after seeing all the red tape
> I'd have to cut through (i mean: Job Title? Number of Employees? Hey, I
> started entering bogus stuff, but then decided I'm too old for that).
>
> And those are marked as REQUIRED fields.
>
> Am I mis-interpreting SourceForge's interface? (very well possible!)
> </rant>
So what ? My SF account knows nothing more about me than my e-mail
address, in order to redirect messages to us...@sf.net. I'm not barking
about privacy violations for that.
Anonymous bugreports were once possible, but the amount of spam made
them impractical.
Sending messages to the mailing list is not good either, because
there's no archive and it's not a bug tracking system -- you rely on
somebody else to actually use the tracker.
Conclusion: put up (with SF), or bear with the bug.
-Alex
Andreas Leitgeb
> On 21 jan, 13:25, tomas <to...@floh.bas23> wrote:
>> the honour goes to John Roll [JBR], who describes the bug in
>> <http://wiki.tcl.tk/15512>. How do I do that?
>> It's a very small ommission: in line 116 of grammar_peg/peg_interp.tcl:
>> ict_match_token "Expected $ch"
>> should read
>> ict_match_token $ch "Expected $ch"
>> and yes, it's still in tcllib-1.13
> Sending messages to the mailing list is not good either, because
> there's no archive and it's not a bug tracking system -- you rely on
> somebody else to actually use the tracker.
Perhaps that was driven by the hope, that on the mailing-list, some
developer would read it, and instantly fix it (based on that the patch
is so trivial), bypassing ticketing altogether for this one.
Evil Son
> ...
> There's something called OpenID, whereby you use an already
> established login from Yahoo, Google, AOL (do they even still
> exist??), LifeJournal, Flickr, Technorati, Wordpress, Blogger, Vidop,
> Verisign, ClaimID, OpenID, or MyOpenId and use that to log into the
> system.
>
> If you happen to have one of those ids, then you should be able to go
> the openid route.
I've just tried the OpenID approach suggested above and found it very
easy.
Evil Son
Alexandre Ferrieux
wrote:
>
> Perhaps that was driven by the hope, that on the mailing-list, some
> developer would read it, and instantly fix it (based on that the patch
> is so trivial), bypassing ticketing altogether for this one.
... which is exactly the kind of counter-productive interaction that
trackers were designed to avoid !
Counter-productive, because (1) all the other users who may encounter
the bug (and not upgrade immediately, eg because they only upgrade to
packaged versions) will rediscover it independently, and (2) because
the fixing commit will lack much of the interesting context, making it
hard for other maintainers to reuse that knowledge.
-Alex
Andreas Leitgeb
> On 21 jan, 15:29, Andreas Leitgeb <a...@gamma.logic.tuwien.ac.at>
> wrote:
>> Perhaps that was driven by the hope, that on the mailing-list, some
>> developer would read it, and instantly fix it
>
> trackers were designed to avoid !
>
> Counter-productive, because (1) all the other users who may encounter
> the bug (and not upgrade immediately, eg because they only upgrade to
> packaged versions) will rediscover it independently, and (2) because
> the fixing commit will lack much of the interesting context, making it
> hard for other maintainers to reuse that knowledge.
In principle, you're right, but if the maintainer had found the bug
himself, I doubt that he'd have created a ticket for it, instead of
just fixing it. But then again, I might be wrong, and he would have
created a ticker for it...
PS: anything new in the "Re: Ideas toward the GC"-thread?
tomas
Thanks all for your comments. I ended up registering an account with SF
(I'm now Igor Stravinski, CEO of a 100-people firm ;-) and filing the
bug.
I'm the only one thinking that web interfaces suck much more than they
should?
Regards
-- tomás
Larry W. Virden
> I'm the only one thinking that web interfaces suck much more than they
> should?
No. I wonder whether it is the web-ness of the interface (limited in
the amount of functionality that one can use to be cross-browser
compatible) or if the REAL problem is that given there are so MANY web
pages, developers just are not writing smart UIs (and perhaps the
issue is that they don't know HOW to write a smarter less sucky UI).
tomas
> On Jan 21, 1:46 pm, tomas <to...@floh.bas23> wrote:
>> I'm the only one thinking that web interfaces suck much more than they
>> should?
>
> No.
Hah. Thanks. I was starting to feel strange ;-)
> I wonder whether it is the web-ness of the interface (limited in
> the amount of functionality that one can use to be cross-browser
> compatible) or if the REAL problem is that given there are so MANY web
> pages, developers just are not writing smart UIs (and perhaps the
> issue is that they don't know HOW to write a smarter less sucky UI).
I would say it's the latter. For the record (and to try to sverve back
to topic, sorry for my rather off-topic rant), very simple interfaces,
like wiki.tcl.tk make me feel at home. SourceForge made me downright
sick.
Ah, and back to the original topic: the bugtacker URL is:
<https://sourceforge.net/tracker/?func=detail&atid=112883&aid=3163541&group_id=12883>
It goes to Andreas Kupries, so I'm rather confident that it'll get
resolved in due time.
Regards
-- tomás