{"TODAY", 3, 7},
{"TOMORROW", 3, 8},
}
time(&thetime);
mytime = localtime(&thetime);
if (NewExtWDay == 7)
NewExtWDay = mytime->tm_wday;
{
}
if (NewExtWDay == 8)
NewExtWDay = mytime->tm_wday + 1;
{
I am not completely happy with how the wild card operates. It
probably sorts the files by their suffix and uses the highest number.
How this works out at the changeover from one year to the next may be implemented in makenl, but I am not shure.
It would suit me better if files were sorted by creation date like in
"ls -t ". If such a construct would work for all the versions that exits is questionable.
@MSGID: 2:280/5003.4 5a172ead
@CHRS: LATIN-1 4
@TZUTC: 0100
@TID: hpt/lnx 1.4.0-sta 16-02-06
Hello Andrew!
In the last few years it has come in vogue create daily nodelist.
In order to accomplish this, the PUBLish <DAY> has to be changed
every day. Although it is not difficult o use some form of scripting
to accomplish this, it would be more convenient if "PUBLish" TODAY
could be used in the control file. For those who want to make the file
one day in advance (before 24:00) "TOMORROW" would also be a usefull publishing day.
I have had a look at the code, and applied the following patch:
---8<----------------------------------------------------------------- ------- diff -r makenl_ng/makenl-code/src/config.c makenl_fg/makenl-code/src/config.c 399a400,401
{"TODAY", 3, 7},
{"TOMORROW", 3, 8},
798a801,810
}
time(&thetime);
mytime = localtime(&thetime);
if (NewExtWDay == 7)
NewExtWDay = mytime->tm_wday;
{
}
if (NewExtWDay == 8)
NewExtWDay = mytime->tm_wday + 1;
{
---8<----------------------------------------------------------------- --------
When using this patch, it is of importance, that
in the files section of the controlfile wilcards are used.
Region 28 region28.* 2:28/0
If the wildcard is omitted, only files of the same publishing day will
be looked for. When not found, older files published on the same
weeday wil be used. (There is aworkaround, but that is not relevant
here).
I am not completely happy with how the wild card operates. It probably sorts the files by their suffix and uses the highest number.
How this works out at the changeover from one year to the next may be implemented in makenl, but I am not shure.
It would suit me better if files were sorted by creation date like in
"ls -t ". If such a construct would work for all the versions that exits is questionable.
I hope that the main users would like the above patch to be included
in the mainstream version of makenl_ng. When the keywords TODAY and TOMORROW are not used, the code is just idle.
The most reliable way would be, to read the first line of the file and
try to determine the date from that, before using the file date. (file dates can change)
In the last few years it has come in vogue create daily nodelist.
In order to accomplish this, the PUBLish <DAY> has to be changed every
day.
Although it is not difficult o use some form of scripting to accomplish this, it would be more convenient if "PUBLish" TODAY could be used in
the control file. For those who want to make the file one day in advance (before 24:00) "TOMORROW" would also be a usefull publishing day.
There are 4 people in the nodelist concerned by this. 2 are producing
a daily nodelist, maybe 3 but I'm not certain about the 3rd one.
None of these, I think, are in need of such an update. At least I will
not use it most likely.
So I think you are proposing an update which may or may not be used by
a grand total of 1 person.
Worth the trouble?
dayThe most reliable way would be, to read the first line of the file
and try to determine the date from that, before using the file date.
(file dates can change)
This is only the case if the lower level submits updates using MakeNL or other nodelist generating software. For ZCs receiving updates from RCs, this is usually the case. However, it is fairly common for NCs of smaller nets to just edit their segment by hand with a text editor. In cases such as these, you can't expect the segment date to be in any particular format on the first line of the file; in fact it may not even be there at all. Even the default date line inserted by MakeNL isn't in the most convenient format for sorting by date, unless you are going to rely on the 3 digit
number, which will cause issues when crossing year boundaries.
In most cases where a segment is sent shortly after being generated,
the file date stored on disk should be close enough.
So I think you are proposing an update which may or may not be used by
a grand total of 1 person.
Worth the trouble?
There are 4 people in the nodelist concerned by this. 2 are producing a daily nodelist, maybe 3 but I'm not certain about the 3rd one.
None of these, I think, are in need of such an update. At least I will not use it most likely.
So I think you are proposing an update which may or may not be used by a grand total of 1 person.
Worth the trouble?
Besides, the coding and testing time for this change will be minimal, as the code is basically already written.
The most reliable way would be, to read the first line of the file and try to determine the date from that, before using the file date. (file dates can change)
The most reliable way would be, to read the first line of the file
and try to determine the date from that, before using the file date.
(file dates can change)
That is not what you want as same number is in file suffix. But
acually there is only one case where it is needed.
From one source two updates are received on Wednesday, one with the Wednesday daynumber and one with the Friday daynumber. If succesive updates are available, on Thursdays the Friday daynumber will be preferred when the sort is made on suffix. Sorting on c-date would
solve this.
With respect to makenl_ng you have always been hard to convince. ;)
Worth the trouble?
MakeNL is used in other networks besides FidoNet. Several of them do produce daily nodelists.
Worth the trouble?
Yes, I think it is.
First: it is not all that much trouble.
Second: when the option is there, more nodelist clerks will probably
start using it.
Having the day number
in the header and the file extension reflect the number of the day it is actually produced is more logical and transparant.
That depends on the c-date of the file being correct. Which doesn't have
to be the case. That's why I suggested to use the full "internal" date in the file.
From one source two updates are received on Wednesday, one with the Wednesday
daynumber and one with the Friday daynumber. If succesive updates are available, on Thursdays the Friday daynumber will be preferred when the sort is made on suffix. Sorting on c-date would solve this. For the odd occation
where the above is the case, it is probably not worth the effort to
have it implemented. I think Ward will agree. ;)
That depends on the c-date of the file being correct. Which doesn't
have to be the case. That's why I suggested to use the full
"internal" date in the file.
And what about if it passed through ERRFLAGS ?
With respect to makenl_ng you have always been hard to convince. ;)
But you cannot claim that in the end I do not fall for proper argumentation.
In the DAILYLST story Z2 led the pack.
To me it looks like a techie-thing ... "Nothing techie has been done for a while, let's do something techie...
gives the FTSC something to document"
where the above is the case, it is probably not worth the effort to
have it implemented. I think Ward will agree. ;)
Do I have to? 8-)
The point I would like to make is that it is absolutely silly to
produce 2
ZONE1 files on Wednesday with an interval of 2-3 hours, one having the Friday Julian-date.
There will be another ZONE1-file a day later which may contain more current information than the ZONE1-file with Friday's Julian date ... hence on Friday a ZONE1-file with non-current information risks to be used when sticking to Wednesday's file.
All that can be avoided by simply producing a single ZONE1-file every day of the week using a single procedure with a single set of region-files and drop the Wednesday's version with Friday's Julian date.
And what about if it passed through ERRFLAGS ?
What about it?
But you cannot claim that in the end I do not fall for proper
argumentation.
Well for your switch to makenl_ng it took some 12 years.
I agree with you, but I do not meddle in the affairs of wizzards.
It plays with the top-line.I can't see (in the sourcecode) with the topline.
\%/@rdBye/2 Torsten
First: it is not all that much trouble.
And I'm not advocating "against", only highlighting there's no demand.
[Errflags]
It plays with the top-line.I can't see (in the sourcecode) with the topline.
What do you ment in detail?
You seem to be a poor judge when it comes to evaluating future needs...
All the tools are present to achieve exactly that already ... have
been doing that for over 2 decades now.
If it's configured, correct.[Errflags]
It plays with the top-line.I can't see (in the sourcecode) with the topline.
What do you ment in detail?
At some point I asked Johan Zwiekhorst that routines to:
1) Force a Region-line with a CRC-cqlculqtion in it when it is missing
2) Check the CRC and if it were wrong, flag it and re-calculate
Eventually that evolved into recalculating and inserting the CRCthe crc will be generated, if it's configuered. If generating of the crc is configured, errflags will but it inside on top of the submitted file.
"whatever"
If you can't find that, then it means someone tinkered with it.Well, I found the lines. :-)
For the record, I will only work with ERRFLAGS-versions produced byI know. I'm getting the information pages of your errflags 2.15. My Downlinks getting the information pages from my errflags 2.18 /OS2 EcomStation 2.2 compiled on my lokal machine on 29.10.2017 0:00h
Niels Joncheere (2:292/0) and Tom De Puysseleer (2:29/0).
\%/@rdBye/2 Torsten
Stephen King "Rita Hayworth and the Shawshank Redemption") and the subsequent abandonement by net-292 who pulled out of your region-28, re-establishing region-29, because it was unanymously decided by all
nodes and its NC you were impossible to deal with.
Well, I found the lines. :-)
Because of your message about errflags, I've had a nice evening with a quite old pascal source-codes.
Tha for that. :-) And now I'm getting a propper cold beer.
All the tools are present to achieve exactly that already ... have Mv>WD> been doing that for over 2 decades now.
23 years by your own count.
Another 16 years and tou will have toppe Rubert Mugabe.
Gna, to spent the time it was nice and usefull. I got quite more involved into the code. ;-)Well, I found the lines. :-)At least I hope I haven't spoiled too much of your time.
Because of your message about errflags, I've had a nice evening
with a quite old pascal source-codes. Tha for that. :-) And now
I'm getting a propper cold beer.
Cold beer ... good idea.;-)
Have a nice week-end ... I will, lotsa non-fido stuff coming up ... including beer tomorrow ... and Sunday too.Have fun.
\%/@rdBye/2 Torsten
Next
spring we've a lot open air music events with some international bands
and starlets. Also some very intresting metal-music events with bands of norway, denmark, sweden, germany and switzerland.
And what about if it passed through ERRFLAGS ?
What about it?
It plays with the top-line.
I have had a look at the code, and applied the following patch:
Region 28 region28.* 2:28/0
It would suit me better if files were sorted by creation date like in
"ls -t ". If such a construct would work for all the versions that exits is questionable.
I'd like to help you test this fix but I do not understand how to use binary executable diffs. I'm currently using MakeNL 3.4.6 (Linux). Could you please drop a copy of your patched Linux executable to me? Pretty please?
Trust me on this... the wildcarding does not matter when there is only
a
single segment available. :)
The are the some values you fave to edit in the first line of the produced Nodelist. The date and the daynumber. Changing this line has
no influence on the CRC. Ofcourse the daynumber suffix has to be
changed as well.
With some tinkering this can all be done with a batch file.
With some tinkering this can all be done with a batch file.
I use this command sometimes to create a first line sometimes when updating segments for the AmigaNet nodelist:
date +';A Region Nodelist for %A, %B %-d, %Y -- Day number %j : 00000'
I just love timely replies ... 8-)
Date: 26 Sep 18 21:31:39 -------------------->-----------------+
From: Paul Quinn !
To: Kees van Eeten !
Subj: Dailylists with makenl_ng ! ________________________________________________________________!
!
Hi! Kees, !
!
On 23 Nov 17 20:52, you wrote to Andrew Leary: !
========= ------------<------------<----------------<--------+
8-)
Anyway if you run into troubles, let me know, I have tried many
variations, but solutions vary on what list you are producing.
I use this command sometimes to create a first line sometimes when updating segments for the AmigaNet nodelist:
date +';A Region Nodelist for %A, %B %-d, %Y -- Day number %j : 00000'
Might get you started...
Sysop: | DaiTengu |
---|---|
Location: | Appleton, WI |
Users: | 910 |
Nodes: | 10 (0 / 10) |
Uptime: | 252:09:55 |
Calls: | 12,118 |
Calls today: | 1 |
Files: | 186,506 |
Messages: | 2,227,583 |