Subject: | Re: Capture Server Time ?
| Date: | Thu, 8 Feb 2024 12:17:29 -0800
| From: | Peter <peterspammenot@whiteknight.email>
| Newsgroups: | pnews.paradox-programming
|
Textstream is twice as fast, slowest 16ms, fastest 15ms.
I had to put in a 1ms sleep() because in a loop of 20 passes I got only
two millisec values. Problem solved.
Thanks again Liz.
On 02/07/2024 10:09 p.m., Peter wrote:
> Good idea, I'll try the textstream tomorrow.
>
> The code is in a library. Every form opens that particular library. If
> nothing is happening then the lock file's timestamp will not be updated.
> I have to be sure that the calling form is getting a current time stamp,
> hence the tcursor lock to trigger the lck file's timestamp update.
>
> Thanks Liz
>
>
> On 02/07/2024 09:29 p.m., Liz McGuire wrote:
>> You could try a textstream type and see if that's any faster. Just
>> read the time off the txt file rather than the lock file. But if all
>> instances are updating the same files, couldn't you just read the lock
>> file without the tcursor? Guess I'm not sure what exactly is going on...
>>
>> Liz
>>
>>
>> On 07 Feb 2024 13:07, Peter wrote:
>>> Back at this again.
>>> I could not find an api or anything else so I decided to pursue
>>> trying to use properties of a file on the server. All I need to do is
>>> get a Date/timestamp for when a transaction is done. Accuracy is not
>>> that crucial. Here is what I came up with.
>>>
>>> method pushButton(var eventInfo Event)
>>> var
>>> tcu tcurosr
>>> dattim datetime
>>> fsy filesystem
>>> endVar
>>>
>>> ;//ServerTime.db is dedicated tbl for this method
>>> ;//crucial that ServerTime.db is located on the server
>>>
>>> tcu.open(":LOBBY:ServerTime")
>>> tcu.lock("read")
>>>
>>> ;//PdoxUsrs.lck is created when lock created
>>> if fsy.findFirst(":LOBBY:PdoxUsrs.lck") then
>>> ;//represents datetime of server
>>> DatTim = fsy.time()
>>> endIf
>>>
>>> tcu.unLock("read")
>>> tcu.close()
>>>
>>> endMethod
>>>
>>> Tested this in rapid-fire using for/endfor loop and writing to array,
>>> including millisecs. The results show a range of 30ms-50ms for twenty
>>> passes.
>>>
>>> The question, finally, is there anything faster I can do than a
>>> record lock to trigger a change in PdoxUsrs.lck? Table lock seems to
>>> be slower. Thank you.
>>>
>>>
>>> Peter
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> On 06/09/2022 12:47 p.m., Peter wrote:
>>>> Is it possible to record events using a server's time rather than
>>>> individual stations? There are two instances, one running an actual
>>>> MS server, the other peer to peer.
>>>>
>>>> The stations cannot be synchronized to the server nor the internet
>>>> so I am running into problems in recording times for certain
>>>> transactions such as "updated xx minutes ago"
>>>>
>>>> The only workaround I can think of is to write a file on the
>>>> 'server' and then reading the timestamp.
>>>>
>>>>
>>>> Peter
|