Re: patch, problems, question
Shannon Weyrick <sweyrick-UWl7u1r57vBWk0Htik3J/[email protected]> Fri, 13 May 2005 10:27:55 -0400
| Newsgroups | gmane.text.rlib.user |
|---|---|
| Message-ID | <[email protected]> |
Bob Doan wrote: > Hi > > >>I've got a patch for html.c that adds a "suppress_head" parameter to not >>output the <body> and <head> tags. It's in my cvs tree - is there a way >>to commit it or should I send a patch to the list (or you)? > > > Unified diff please. > > Can you send it to rlib-devel-TtF/[email protected] > Ok, sent. > > >>Secondly - I've been having a segfault problem lately that started a few >>weeks ago (from CVS build). What happens is when rlib has an error (say >>a BAD_OPERAND or bad SQL) it will segfault, the apache child will die, >>and it won't print out an error message - I just see a message in apache >>error log that the child died. However it only happens after it's been >>running for awhile - if I immediately restart the server it will run ok >>and rlib will show the error message. If it's been running for more than >>5 or 10 minutes say, that's when it crashes. >> >>I wish I had time to track it down more than that, but I haven't at the >>moment. I realize it will be hard to look into unless it can be >>duplicated. It's more if a heads up and to see if anyone else is >>experiencing this. > > > Hym... Dunno. We have The lasted RLIB in a production environment and > it's not crashing. > Are you using it with PHP 4.3.x under Apache 1.3.x? When I get a chance I'm going to run it under gdb and see where the segfault lies - or at least find a way to consistently reproduce it. > >>Thirdly, I'm interesting in adding support to the PHP binding to lookup >>a variable in a hash (in addition to a single variable, as now). I >>figure this can be done with another attribute to the <field> tag - >>"index" for example. I looked into it a bit and I can add this much to >>php/environment.c if rlib_php_resolve_memory_variable gets two >>parameters. I can see where to add it to the structure to keep the >>"index" string (in parsexml.c). The tough part is pcode.c -- how can I >>get access to that index string in rlib_new_operand? >> >>Or, what's the better way to implement it? > > > Well.. Hym...... In the long term it would be good if RLIB could hash > and haves lists(arrays) of data for other purposes. > > This would be kinda hard. > > What would be easier to be able to pass RLIB custom functions (even from > binding languages like PHP) > > So some new api like > rlib_add_function(r, "custom_function", pointer_to_function) > > Then in the xml you could: > <field value="custom_function('ht_name', 'ht_value')"> > > In order to do this we would have to also make public rlib's stack > (push, pop) > > In php you would do: > > function custom_function() { > $ht_value = rlib_SPECIAL_POP($rlib); > $ht_name = rlib_SPECIAL_POP($rlib); > $value = $$ht_name[$ht_value]; > $ht_name = rlib_SPECIAL_PUSH($rlib); > } > > > Freaky eh? > Definitely. The ability to hook into a custom function would be great. The PHP interface seems a little low level, though, I wonder if there's a way to hide some details. If it gets that far, it might not be too much of a stretch to run built in PHP functions on rlib data. That would open up a great wealth of functions, obviously. In the short term however, can you think of a way to get rlib_php_resolve_memory_variable access to an optional index string so that if Z_TYPE_PP(data) == IS_ARRAY it can look it up? Thanks, Shannon ------------------------------------------------------- This SF.Net email is sponsored by Oracle Space Sweepstakes Want to be the first software developer in space? Enter now for the Oracle Space Sweepstakes! http://ads.osdn.com/?ad_id=7393&alloc_id=16281&op=click