Hi Dmitry,
>> May be solution is to decrease key length - for example use 16 byte
>> md5 value, instead of 32 chars value. In any case they are equal.
>> Binary comparison is always faster then text comparison.
DD> No. 16 bit md5 is not unique. It is wrong to truncate md5 even to 31 bytes.
You d
Hi!
Dmitry Martynenko wrote:
> May be solution is to decrease key length - for example use 16 byte
> md5 value, instead of 32 chars value. In any case they are equal.
> Binary comparison is always faster then text comparison.
No. 16 bit md5 is not unique. It is wrong to truncate md5 even to 31 by
Hi Dmitry,
DD> I do not remember why SQL_NO_CACHE is there. I remember there was
DD> a reason for it. You can try to remove that and see if it helps.
Don't worry about it.
Such SQL queries are slow with and without it (by my tests).
I think you use SQL_NO_CACHE to avoid filling mysql cache with
Hi!
Dmitry Martynenko wrote:
> Today I have tried to understand RealURL algorithm when it decode
> urls...
> My brain begun to boil :)
Welcome to the RealURL world! :D
> But two things I have noticed:
>
> 1) tx_realurl_chashcache table has duplicates in chash
> spurl_hash
Hi Dmitry,
>> But after 3 days of site work USER_INT links with cHash appears again!
>> They have small percent for now - only about 0.5-1%, but they exists.
DD> Can you compare MD5s for these URLs?
When I have some free time I investigate this problem deeply..
Today I have tried to understand
Hi!
Dmitry Martynenko wrote:
> But after 3 days of site work USER_INT links with cHash appears again!
> They have small percent for now - only about 0.5-1%, but they exists.
Can you compare MD5s for these URLs?
--
Dmitry Dulepov
TYPO3 core team
In the blog:
http://typo3bloke.net/post-details/h
Hi Dmitry,
>> DM> PS: I'll try to implement localy patch from
>> DM> http://bugs.typo3.org/view.php?id=5237
>>
>> This patch does not work for us.
DD> I think there should me a full md5. hexdec on md5 comes from lack
DD> of understanding that md5 is valid only in full, not in parts. The table
s
Hi Dmitry,
DD> I think there should me a full md5. hexdec on md5 comes from lack
DD> of understanding that md5 is valid only in full, not in parts. The table
should be changed.
DD> I do not fix RealURL at the moment (though I will release a small
DD> update soon). The fix can come only next year
Hi!
Dmitry Martynenko wrote:
> DM> I have added bug recently: http://bugs.typo3.org/view.php?id=9619
> DM> Today I add note from my previous mail.
> DM> It is not so urgent, because only goes to small performance decline
> DM> on our server.
>
> It goes to small performance decline, but it also f
Hi Dmitry,
Old thread, but it disturbs us again.
>>> Just let you know.
>>>
>>> After several day when tx_realurl_chashcache fills with few thousand
>>> records the bug with unwanted cHash appears again.
DD>> The information is not useful in any way, sorry. If you wrote bug
DD>> report, it is e
Hi Dmitry,
>> Just let you know.
>>
>> After several day when tx_realurl_chashcache fills with few thousand
>> records the bug with unwanted cHash appears again.
DD> The information is not useful in any way, sorry. If you wrote bug
DD> report, it is enough :)
DD> If you need urgent solution, yo
Hi!
Dmitry Martynenko wrote:
> Just let you know.
>
> After several day when tx_realurl_chashcache fills with few thousand
> records the bug with unwanted cHash appears again.
The information is not useful in any way, sorry. If you wrote bug
report, it is enough :)
If you need urgent solution,
Hi Dmitry,
>> May be you can to do you comment in decodeSpURL_cHashCache()
>> // TODO Dima: should be changed to normal md5 too?
>>
>> And also add clearing of tx_realurl_chashcache table to
>> clearPageCacheMgm() function?
DD> Can you open a bug report and describe it all there? Set severity to
Hi!
Dmitry Martynenko wrote:
> To Dmitry Dulepov :)
>
> May be you can to do you comment in decodeSpURL_cHashCache()
> // TODO Dima: should be changed to normal md5 too?
>
> And also add clearing of tx_realurl_chashcache table to
> clearPageCacheMgm() function?
Can you open a bug report and des
Hi Dmitry,
I have look into TYPO3 Core and RealURL code and investigated
following things.
cHash can be added to URLs in two places:
1. In tslib_cObj class function typoLink()
2. In RealURL extension in function decodeSpURL_doDecode()
I logged typoLink() and it is working fine!
Then I looked in
Hi Dmitry,
>> Links are generated from *one place* in plugin using this function call:
>>
>> $this->singlePageId = $GLOBALS["TSFE"]->id;
>> $linkText = 'link text...';
>> $urlParameters = array('unid' => $rowData['code'], 'photo' => 1);
>> $link = $this->pi_linkTP($linkText, $urlParameters, 1, $t
Hi Dmitry,
>> Links are generated from *one place* in plugin using this function call:
>>
>> $this->singlePageId = $GLOBALS["TSFE"]->id;
>> $linkText = 'link text...';
>> $urlParameters = array('unid' => $rowData['code'], 'photo' => 1);
>> $link = $this->pi_linkTP($linkText, $urlParameters, 1, $t
Hi Dmitry,
>> Links are generated from *one place* in plugin using this function call:
>>
>> $this->singlePageId = $GLOBALS["TSFE"]->id;
>> $linkText = 'link text...';
>> $urlParameters = array('unid' => $rowData['code'], 'photo' => 1);
>> $link = $this->pi_linkTP($linkText, $urlParameters, 1, $t
Hi Dmitry,
>> Links are generated from *one place* in plugin using this function call:
>>
>> $this->singlePageId = $GLOBALS["TSFE"]->id;
>> $linkText = 'link text...';
>> $urlParameters = array('unid' => $rowData['code'], 'photo' => 1);
>> $link = $this->pi_linkTP($linkText, $urlParameters, 1, $t
Hi Dmitry,
>> Links are generated from *one place* in plugin using this function call:
>>
>> $this->singlePageId = $GLOBALS["TSFE"]->id;
>> $linkText = 'link text...';
>> $urlParameters = array('unid' => $rowData['code'], 'photo' => 1);
>> $link = $this->pi_linkTP($linkText, $urlParameters, 1, $t
Hi!
Dmitry Martynenko wrote:
> Links are generated from *one place* in plugin using this function call:
>
> $this->singlePageId = $GLOBALS["TSFE"]->id;
> $linkText = 'link text...';
> $urlParameters = array('unid' => $rowData['code'], 'photo' => 1);
> $link = $this->pi_linkTP($linkText, $urlParam
Hi List,
We have own USER_INT plugin, which renreds List and Single views
on one page depending of GET parameters.
I have noticed that urls to Single view generated 50%/50%
with cHash added. Thus it leads to great number of cached variants
of same page contents with USER_INT marker.
For example
22 matches
Mail list logo