Re: 1 cset was pushed to core/sqlite: - replacing a & with @ to suppress the e...
Marcel van der Boom <[email protected]>
| Newsgroups | gmane.comp.cms.xaraya.bk-notices |
|---|---|
| Organization | Xaraya |
| Message-ID | <[email protected]> |
Agreed, that was the case initially. However, that makes things unnecessary hard. The error hiding is only warranted when a statement is made deliberately to fail and check for the outcome and react to that. In all other cases it just hides the error. That has been a bug for a long time and it took me a while to sort out why xar_pages didnt install, because you just go looking in the wrong direction, because the error displayed is merely a consequence of the real error. I made a note in the code to the extent of: "We dont wanna do this here, but at the driver level, but we cant right now" marcel Jason Judge wrote: > Looking at the default ADOdb execute method, it does suppress its own PHP > errors using '@'. I think therefore that the sqlite driver needs to do the > same thing, rather than at the higher level as you have done here. > Basically, any driver raising a PHP error or warning is considered fatal. > They should suppress those errors, and capture them separately, then pass > them to the ADOdb error handler, which in turn passes them to Xaraya. I've > pushed a change to the data dictionary (adodbupdate scenario) to suppress > the raising of the error within ADOdb, but it is up to the driver to > suppress the PHP error. > > -- JJ > > > "Jason Judge" <[email protected]> wrote in message > news:[email protected]... > >>I've raised this with ADOdb. I suspect it is an issue with the drivers, >>rather than the data dictionary, since some drivers will raise a PHP > > warning > >>if the table does not exist, and some will not. It all depends on what SQL >>statement the meta columns fetch is executing, so the warnings should be >>suppressed there IMO. >> >>-- JJ >> >> >>http://phplens.com/lens/lensforum/msgs.php?id=11956 >> >>"Marcel van der Boom" <marcel-111RdWWlOxkOkXtL6OYfHgC/[email protected]> wrote in message >>news:[email protected]... >> >>>Changes were applied by marcel to core/sqlite level 15. >>> WARNING: These changes bypassed the Quality Assurance Tests! >>> The following cset was applied: >>> >>> ChangeSet >>> 1.2029 05/02/21 14:11:26 [email protected] +2 -0 >>> - replacing a & with @ to suppress the error which will be > > generated > >>if the changeTable method is used to create tables >> >>> (yes, that sound unintuitive, but apparently that is the case). > > This > >>fixes the blocker for installing xarpages with sqlite >> >>> and should not hurt other databases (logic didnt change) >>> I'll let jason reconfirm this for other databases. >>> >>> html/xaradodb/drivers/adodb-sqlite.inc.php >>> 1.9 05/02/21 14:11:24 [email protected] +3 -1 >>> Inserting a comment related to the change in the datadict file. >>> >>> html/xaradodb/adodb-datadict.inc.php >>> 1.9 05/02/21 14:11:23 [email protected] +6 -1 >>> - replacing a & with @ to suppress the error which will be >> >>generated if the changeTable method is used to create tables >> >>> (yes, that sound unintuitive, but apparently that is the case). >> >>This fixes the blocker for installing xarpages with sqlite >> >>> and should not hurt other databases (logic didnt change) >>> I'll let jason reconfirm this for other databases. >>> >>>Links for the differences: >>>http://xaraya.com:15000/core/sqlite/[email protected] >> >> > >