IMAP: Group xxx is not uidvalid: what to do?
Dryice Liu
I'm using Gnus 5.8.8.
I encounter a "not uid-valid" error when I tried to enter the Inbox
folder of my IMAP server after I move (B m) all the articles from the
Inbox folder to other folders. And I can still enter the Inbox folder
after I answer "yes" several times. Other folders of the same account
on the same server don't have the same problem.
I cheched files imap.el, nnimap.el, and ChangeLog of Gnus. And as I
don't know elisp, I can't get too much from the code. And in nnimap.el
the author said "Do The Right Thing when UIDVALIDITY changes (what's
the right thing?)" in the minor todo list, which disappointed me a
little :(
And in the ChangeLog, Simon Josefsson (the author of nnimap.el" said:
2000-10-08 Simon Josefsson <si...@josefsson.org>
* nnimap.el (nnimap-verify-uidvalidity): Delete overview file when
uid validity changes.
(nnimap-group-overview-filename): Store uidvalidity in filenames.
Rename old files into new format.
It seams there is an "overview file" that stores my old uid
validity. And the program tried to delete it but failed on my box. I
guess I can fix the problem if I delete that file manually -- of
cause, I must find it first :)
Can anyone kindly tell me what's the problem here and how can I fix
it? Thanks.
--
Thanks,
Dryice Liu
Kai Großjohann
> I encounter a "not uid-valid" error when I tried to enter the Inbox
> folder of my IMAP server after I move (B m) all the articles from the
> Inbox folder to other folders.
I think Simon has removed that message from recent Oort Gnusae.
In the meantime, M-c (gnus-group-clear-data) on that group might help.
kai
--
GNU/Linux provides a nice `poweroff' command, but where is `poweron'?
Dryice Liu
> Dryice Liu <dryice...@yahoo.com.cn> writes:
>
> > I encounter a "not uid-valid" error when I tried to enter the Inbox
> > folder of my IMAP server after I move (B m) all the articles from the
> > Inbox folder to other folders.
>
> I think Simon has removed that message from recent Oort Gnusae.
>
> In the meantime, M-c (gnus-group-clear-data) on that group might help.
>
Thanks, Kai. gnus-group-clear-data doesn't work. And something I did
similiar to it works:
1. kill group "Inbox"
2. quit Gnus
3. remove the overview directory of "inbox" under
~/News/overview/nnimap/server
4. restart Gnus and re-subscribe group "inbox"
I noticed that the name of the overview directory is the same as the
"uidvilidity" field in ~/.newsrc.eld. And I guess if I can get the new
uidvilidity number, I can change it in ~/.newsrc.eld and move the
original directory to another named by the number, and the problem
should disappear.
And I also guess the problem occured because after I deleted all the
articles in "inbox" folder, the server delete the folder. And when new
message comes, the server create another one.
--
Thanks,
Dryice Liu
Dryice Liu
> Dryice Liu <dryice...@yahoo.com.cn> writes:
>
> > I encounter a "not uid-valid" error when I tried to enter the Inbox
> > folder of my IMAP server after I move (B m) all the articles from the
> > Inbox folder to other folders.
>
> I think Simon has removed that message from recent Oort Gnusae.
>
> In the meantime, M-c (gnus-group-clear-data) on that group might help.
>
Thanks, Kai. gnus-group-clear-data doesn't work. And something I did
similiar to it works:
1. kill group "Inbox"
2. quit Gnus
3. remove the overview directory of "inbox" under
~/News/overview/nnimap/server
4. restart Gnus and re-subscribe group "inbox"
I noticed that the name of the overview directory is the same as the
"uidvilidity" field in ~/.newsrc.eld. And I guess if I can get the new
uidvilidity number, I can change it in ~/.newsrc.eld and move the
original directory to another named by the number, and the problem
should disappear.
And I also guess the problem occured because after I deleted all the
articles in "inbox" folder, the server delete the folder. And when new
message comes, the server create another one. So, I guess I can avoid
the same problem by never empty any folder -- until I switch to a
newer version of Gnus :)
--
Thanks,
Dryice Liu
Kai Großjohann
> I noticed that the name of the overview directory is the same as the
> "uidvilidity" field in ~/.newsrc.eld. And I guess if I can get the new
> uidvilidity number, I can change it in ~/.newsrc.eld and move the
> original directory to another named by the number, and the problem
> should disappear.
No! No!
The file will have the wrong data. That's the whole point of
uidvalidity: each message can be identified by message id or by
number. IMAP does not guarantee the numbers to stay the same always,
but it provides the uidvalidity value. And there is a guarantee: as
long as the uidvalidity value stays the same, the numbers stay the
same, too.
And since overview files use the number to identify the message, using
the same overview file for different uidvalues means that the numbers
in the overview file and the numbers on the server possibly don't
match.
I suggest to just delete the old overview file.
Dryice Liu
> > I noticed that the name of the overview directory is the same as the
> > "uidvilidity" field in ~/.newsrc.eld. And I guess if I can get the new
> > uidvilidity number, I can change it in ~/.newsrc.eld and move the
> > original directory to another named by the number, and the problem
> > should disappear.
>
> No! No!
>
> The file will have the wrong data. That's the whole point of
> uidvalidity: each message can be identified by message id or by
> number. IMAP does not guarantee the numbers to stay the same always,
> but it provides the uidvalidity value. And there is a guarantee: as
> long as the uidvalidity value stays the same, the numbers stay the
> same, too.
>
> And since overview files use the number to identify the message, using
> the same overview file for different uidvalues means that the numbers
> in the overview file and the numbers on the server possibly don't
> match.
Hmmm, Seams I must read IMAP protocal before I talk with you :) And I
guess Gnus can handle this: once I copied an overivew file from
another group, and Gnus correted it :)
>
> I suggest to just delete the old overview file.
Thanks, I've done that.
--
Thanks,
Dryice Liu
Brian P Templeton
[...]
> I think Simon has removed that message from recent Oort Gnusae.
^^^^^^
I don't think that -ae is the correct suffix for Gnus. Assuming Gnus
is a Latin wordš, it is not a first-declension noun, because it does
not end with -a. It must be either third-declension (no pattern for
nominative singular noun endings) or fourth-declension (nominative
singluar nouns end with -us). If it is a third-declension noun, it
could be masculine, femenine, or neuter; if it a fourth-declension
noun, it is either masculine or neuter. So the plural of Gnus should
be either Gnuses (if it is a third-declension noun), or either Gnüs
(notice the long `u') or Gnusua, if it is a fourth-declension noun.
There is not enough information given to tell if it is
third-declension or fourth-declension˛.
[...]
Footnotes:
š We all know the Romans had Emacs. The Roman Empire fell only when
the kingdom split between the GNU Emacs-using Roman Empire to the
west of Italy (and including Italy) and the XEmacs-using Byzantine
to the east... The Roman Empire finally fell after being invaded by
eVIl VI users. (Apparently Nova Roma
(<URL:http://www.novaroma.org/>) is going to use TUNES Emacs ;).)
˛ But my Latin class has not yet officially learned about
fourth-declension nouns, so I may be incorrect.
--
BPT <b...@tunes.org> /"\ ASCII Ribbon Campaign
backronym for Linux: \ / No HTML or RTF in mail
Linux Is Not Unix X No MS-Word in mail
Meme plague ;) ---------> / \ Respect Open Standards
Kai Großjohann
> Kai.Gro...@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
>
> [...]
>> I think Simon has removed that message from recent Oort Gnusae.
> ^^^^^^
> I don't think that -ae is the correct suffix for Gnus.
Lars says so, so it must be true.
Thou shalt not doubt the wisdom of Lars the Magnificent!
Paul Jarc
> š We all know the Romans had Emacs. The Roman Empire fell only when
> the kingdom split between the GNU Emacs-using Roman Empire to the
> west of Italy (and including Italy) and the XEmacs-using Byzantine
> to the east... The Roman Empire finally fell after being invaded by
> eVIl VI users.
Robert Firth has observed that one of the main reasons for the fall of
the Roman Empire was that lacking any representation of 0, they had no
way to indicate successful termination of their C programs. Just
imagine how dreadful it would be if Emacs appeared to fail.
paul
Jonadab the Unsightly One
> Kai.Gro...@CS.Uni-Dortmund.DE (Kai Großjohann) writes:
>
> [...]
> I think Simon has removed that message from recent Oort Gnusae.
> ^^^^^^
> I don't think that -ae is the correct suffix for Gnus. Assuming Gnus
> is a Latin word,
[Incidentally Gnus does not correctly rewrap the above. It messes up
the caret-underscoring. I had to make manual adjustments.]
Etymologically, Gnus is formed from the plural of the English
word gnu, which comes from Hottentot[1], not Latin.
> it is not a first-declension noun, because it does
> not end with -a. It must be either third-declension (no pattern for
> nominative singular noun endings) or fourth-declension (nominative
> singluar nouns end with -us).
If the us were the Latin singular ending, then the plural would be
Gni. However, that would imply that the root is "Gn", and I don't
believe that is the case. The root must surely be at least "Gnu" and
possibly all of "Gnus". And anyway, it clearly doesn't derive from
Latin, at least not anywhere near directly.
> If it is a third-declension noun, it
> could be masculine, femenine, or neuter;
Any gender the original word had was lost when the word "Gnu" was
imported into English, because English nouns generally don't have
gender. Since "Gnus" comes from the English form, it presumably
doesn't have gender, either.
The question of the plural form of Gnus is a complex one, precisely
because the singular comes from what was originally a plural word in
English (albeit a plural word with a rather different meaning). The
plural form was obviously chosen because of its similarity to "news",
which is also related etymologically and has a lot to do with the
meaning (albeit not in the original English sense of the word "news"
but in the sense that has to do with usenet). This raises the
question of the etymology of "news". Interestingly, "news" comes from
the plural of the adjective "new" (used, one presumes, substantivally)
and is related to the Old French[1], for whatever that's worth.
If English has clear rules for forming the plurals of words that come
from the plurals of other words, I don't know what they are. I'll
list examples I know about:
person ==> people ==> peoples (i.e., ethnic groups)
fish ==> fish ==> fishes (i.e., types of fish)
thou ==> you ==> y'all ==> all y'all[2]
The last is a pronoun and thus cannot reasonably be expected to follow
the noun pattern exactly. I'm sure there are additional examples; if
anyone could point out some more, it might help establish a pattern,
or the lack thereof. Perhaps the most important pattern clue we could
gain would be from the way "news" gets pluralised; however,
unfortunately, it generally doesn't, but instead gets applied as an
adjective modifiying a plural noun (news media, newspapers, news
sources, news items, news stories). Gnus, however, clearly needs a
direct plural, to refer to different instances or versions of Gnus.
However, the pattern above (add s) may not apply to Gnus, since Gnus
ends with s already, which has a strong tendency to alter the way a
plural is formed in English. Pronouncing the existing s as z and
adding es seems logical, but I don't think it's a foregone conclusion,
unless a more thoroughgoing survey of the plurals of words that derive
from plurals shows that they form their plurals regularly. I don't
think two nouns constitute enough pattern to say that Gnus must follow
the same pattern. In particular, unless someone can find another word
that comes from a plural form _ending in s_ and forms its plural by
pronouncing that as z and adding es, I would definitately hesitate to
jump to the conclusion that Gnus should form its plural that way.
[1] OED, 1971 Compact Edition.
--
(require 'kill-a-lawyer)
(setq kill-a-lawyer-mode t)
Glyn Millington
>
> If the us were the Latin singular ending, then the plural would be
> Gni.
Hum - well if it were a _fourth_ declension masculine/feminine Latin noun
the plural of Gnus would be, wait for it, Gnus (with a long u). If it
were neuter it would be Gnua (all these nominative of course). As a
first declension noun Gni just sounds _wrong_!
But as you say, Gnu/s/ae has nothing much to do with Latin in its
origins...
felix qui potuit rerum cognoscere causas.........
> (require 'kill-a-lawyer)
> (setq kill-a-lawyer-mode t)
If that isn't a major mode, maybe it should be?
Glyn
--
**************************************************
* Here we are then....... *
* http://members.tripod.co.uk/Christchurch2000uk *
**************************************************
Luis Fernandes
Lars, I was going to suggest making the original post a footnote to
the first occurence of the words "Gnusae" in the Gnus manual but I
think jonadab's followup would make a better footnote.
Paul Jarc
> Brian P Templeton <b...@tunes.org> writes:
>> it is not a first-declension noun, because it does
>> not end with -a. It must be either third-declension (no pattern for
>> nominative singular noun endings) or fourth-declension (nominative
>> singluar nouns end with -us).
>
> If the us were the Latin singular ending, then the plural would be
> Gni.
Only if it were second declension. Fourth is also possible. I can't
think of any third declension nouns ending in -us.
> If English has clear rules for forming the plurals of words that come
> from the plurals of other words, I don't know what they are.
I'd say English is already sufficiently irregular that "Gnusae" will
do just fine (especially since the word was invented anyway).
paul
--o0o--
> ends with s already, which has a strong tendency to alter the way a
> plural is formed in English. Pronouncing the existing s as z and
> adding es seems logical, but I don't think it's a foregone conclusion,
> unless a more thoroughgoing survey of the plurals of words that derive
> from plurals shows that they form their plurals regularly. I don't
> think two nouns constitute enough pattern to say that Gnus must follow
> the same pattern. In particular, unless someone can find another word
> that comes from a plural form _ending in s_ and forms its plural by
> pronouncing that as z and adding es, I would definitately hesitate to
> jump to the conclusion that Gnus should form its plural that way.
Well, perhaps one should consider the plural form of Emacs as a precedent.
Klaus Alfert
> Dryice Liu <dryice...@yahoo.com.cn> writes:
>
> > I encounter a "not uid-valid" error when I tried to enter the Inbox
> > folder of my IMAP server after I move (B m) all the articles from the
> > Inbox folder to other folders.
>
I have had a similar problem, but not in the Inbox, it was in another
folder used as a target for splitting.
> In the meantime, M-c (gnus-group-clear-data) on that group might help.
this didn't work for me. But after deleting the uid-value in the group
parameters (i.e. setting it to nil), it worked.
Klaus.
--
--------------------------------------------------------------------
Klaus Alfert alf...@ls10.de
Software Technology +49+231+755-4812
CS Department, University Dortmund
Jonadab the Unsightly One
> > I would definitately hesitate to jump to the conclusion that Gnus
> > should form its plural that way.
>
> Well, perhaps one should consider the plural form of Emacs as a precedent.
So, you suggest Gnusen?
As fun as that is to pronounce, I'm tempted to agree.
Jonadab the Unsightly One
> > (require 'kill-a-lawyer)
> > (setq kill-a-lawyer-mode t)
>
> If that isn't a major mode, maybe it should be?
I think it is a minor mode. It automagically narrows buffers to
hide the copyright notices at the beginnings of source code.
--
Jonadab the Unsightly One
> "--o0o--" <blahd...@wahwahwah.boohoo> writes:
>
> > > I would definitately hesitate to jump to the conclusion that Gnus
> > > should form its plural that way.
> >
> > Well, perhaps one should consider the plural form of Emacs as a precedent.
>
> So, you suggest Gnusen?
>
> As fun as that is to pronounce, I'm tempted to agree.
Though upon further reflection, Emacs forms its plural that way
not merely because of the s, but at least partly because the cs
combination is pronounced like "x".
Brian P Templeton
forget to suggest that the plural could be Gni (I think I omitted it
because it didn't seem like it possibly be correct), but then it would
be a second-declension noun.