On Thu, 18 Jan 2007, Giuliano Gavazzi wrote:
Actually I don't. Either you or me have misunderstood how buckets work,
since the rest of the code should syntactically be equivalent. Or I'm
missing some fine detail somewhere.
Perhaps I do not understand buckets fully (and brigades), but this
On Wed, 17 Jan 2007, Giuliano Gavazzi wrote:
I have a solution for the r470455 mod_disk_cache not caching SSI.
There are two points where the module seems incorrect to me, changing those
makes it work:
Since you're talking about the code on trunk, you should be warned
that the current state
On 17 Jan 2007, at 10:36, Niklas Edmundsson wrote:
On Wed, 17 Jan 2007, Giuliano Gavazzi wrote:
I have a solution for the r470455 mod_disk_cache not caching SSI.
There are two points where the module seems incorrect to me,
changing those makes it work:
[...]
First, don't reindent code
On 17 Jan 2007, at 12:15, Giuliano Gavazzi wrote:
note that I am still calling store_headers, for it has side effects
(presumably the stuff you say it should do). But store_disk_headers
should not, unless it overwrites the headers when called again with
the correct size. But this is not
On Wed, 17 Jan 2007, Giuliano Gavazzi wrote:
rv = apr_file_seek(dobj-hfd, APR_SET, off);
does not rewind if the file has been opened with APR_FOPEN_BUFFERED. Now, I
This is an APR bug, I submitted a bug report for it a while ago. I
worked around it by not using buffering at all
On 17 Jan 2007, at 16:06, Niklas Edmundsson wrote:
On Wed, 17 Jan 2007, Giuliano Gavazzi wrote:
rv = apr_file_seek(dobj-hfd, APR_SET, off);
does not rewind if the file has been opened with
APR_FOPEN_BUFFERED. Now, I
This is an APR bug, I submitted a bug report for it a while
On Wed, 17 Jan 2007, Giuliano Gavazzi wrote:
Is this in trunk or in 2.2.4 proper? You should probably ignore
mod_disk_cache on trunk until that situation is settled.
I could work on mod_disk_cache from 2.2.4 proper, and find what causes the
bug with SSI pages, but I do not see why I should
On 17 Jan 2007, at 17:32, Niklas Edmundsson wrote:
patches! And I suppose now you agree that the extra indentation on
my patch stemmed from the broken nature of the original code!
Actually I don't. Either you or me have misunderstood how buckets
work, since the rest of the code should
I have a solution for the r470455 mod_disk_cache not caching SSI.
There are two points where the module seems incorrect to me, changing
those makes it work:
1) in store_body the condition (!APR_BUCKET_IS_EOS(APR_BRIGADE_LAST
(bb))) was incorrectly stopping the flow from ever going past (for
Here is the patch against r470455:
mod_disk_cache.c.r470455.patch
Description: Binary data
g
Giuliano Gavazzi wrote:
I have found that mod_disk_cache is definitely not compatible with
mod_include when the SSI pages contain something like !--#exec
cmd=./foo.sh -- (or even an include virtual that points to a cgi).
This isn't so much a cache problem, but rather a problem with the SSI
On Mon, 15 Jan 2007, Graham Leggett wrote:
In order for caches to work, the Last-Modified or ETag headers need to be set
correctly on the content, and this isn't always the case. When this happens,
content isn't cached.
Another module with this problem is the mod_dir directory index
On 15 Jan 2007, at 19:20, Giuliano Gavazzi wrote:
[re r470455 not serving cached content]
This, I think, happens because in mod_cache' s cache_save_filter
you have the following calls is this order:
1) store_headers
2) store_body
but the size of the body is not know until store_body is
On 16 Jan 2007, at 01:08, Giuliano Gavazzi wrote:
r470455
I should have really kept this branch of the thread separate. Let me
just state, to avoid confusion later:
1) r470455 cache imported in the 2.2.4 tag, shows the cache
discarding behaviour when combined with SSI exec cmd. So it
Hello, I am new to this list and hope this will be in an acceptable
format and of appropriate content.
I have found that mod_disk_cache is definitely not compatible with
mod_include when the SSI pages contain something like !--#exec
cmd=./foo.sh -- (or even an include virtual that points
15 matches
Mail list logo