Zeoslib Portal

The ZFormatSettings unit and how it works is very nice. There is an issue, however, because while you can set various formats, you can't set the date or time separators, and so if they get changed from the regional defaults the library routines used for conversions won't work. For example if my regi...

Awesome, thanks for the tips!

-Mark

Hi All! I have a situation where I'm getting a disconnected connection despite doing some attempted keep-alive pinging to the server (in this case with the OLEDB protocol to SQLServer.) What's the recommended way to handle this kind of thing as far as figuring out that it's disconnected and then rec...

In looking at the FormatSettings stuff I noticed that procedure TZParam.SetAsZTimestamps uses LongTimeFormat as the format used to do the conversion to string. I *think* that that should be LongDateFormat instead. I'm not sure if this is a bug or not, it's just something I noticed looking through th...

Hi Michael, Thanks! I see that you've just committed a change to have ZParams use ZFormatSettings if available. Nice! My other comments are just related to Delphi's handling of the 'm' format character. If it appears directly (disregarding separators) after an 'h' character it should be treated as a...

I'm not sure this is useful or not, but what I do to fake a ping to the server is to request a tablelist with a filter that won't exist. I'm not sure it's the way to go, but it's been working pretty well in all my test cases. -Mark if NativePingSupported then begin Result := ZConn.PingServer; end el...

Here's a modification to TryUniToTimeStamp to attempt to handle the 'm' is for minute if it follows 'h' issue. I put a define in just to make it easier to do a/b type testing. I wasn't sure about all the delimiter chars I used to cancel the MIsMinute boolean, so they could probably be adjusted. I'd ...

Thanks Jan! The non-tzfield format handling (ZSysUtils) is used for parameter .value string->date/time/timestamp conversions and that's where I hit some issues. It looks like 'm' or 'mm' for minutes (instead of n) is only handled by the time conversions and not the timestamp conversions. Since the d...

After some more research I realized that I had a fundamental misunderstand of how Zeos handles datetime format settings. I was thinking that the ConSettings's ReadFormatSettings and WriteFormatSettings are set from the ZFormatSettings, but they are not. They're either set by ConnProps_DateReadFormat...

Currently the ZDatasetParams unit uses the older {$IFDEF WITH_FORMATSETTINGS}FormatSettings{$ELSE}SysUtils{$ENDIF}.ShortDateFormat type date/time formattings. I'd like to convert it to use the FConSettings^.ReadFormatSettings.DateFormat type calls instead. However there are a couple of snags: 1. FCo...

[update - moved the ReadFormatSettings stuff to a new post] The new change is good! I just modified the date/time/timestamp conversions to use the nice new ReadFormatSettings. I'm using FConnSettings, but I'm not sure if it should be using a different level. This code was copied from ZDbcCache.pas a...

I did see that one and changed it (yesterday I think.) Does setting the value and then the datatype actually do a conversion? I don't see that in my tests. I see that InternalSetAsXXXX looks like it's capable of doing conversions, but CheckDataIndex sets FSQLDataType to the variant type and then no ...

Hey all! How do you set the value of a ZParam without changing its datatype? Basically I'm wondering if I set the datatype of a ZParam to a date, then how can I set its value using a string while keeping it a datetype? I thought that I could do something like: MyParam.SQLType := stDate; MyParam.Valu...

That's great new! Mystery solved and I'll chill with a brewsky later! I must have misunderstood some of this thread lol. I agree on what you say about the lobs IF the dataset is a one way, firehose type cursor. I'm not sure that's possible in the current architecture (maybe using dbc directly instea...

Those settings are the same as mine (although for my test I'm not threading.) I don't *think* any of my changes made any functional change in the way things work. I did a "one to one" conversion from the old CachedLobs setting to the new LobCacheMode setting. I've gone back to older commit...